 Hi everyone, my name is Louise Kaepner and today I'm going to be leading the Computational Social Science Introductory Workshop. So throughout this workshop we're going to be covering three main sections which include so firstly what is up with computational social science so what does this term even mean then we're going to discuss a bit about how you actually can become a computational social scientist and finally we'll cover the eight steps of CSS which is where we'll have the most interactive part of the workshop and you can sort of get stuck in sketching out your own computational social science projects and of course at the end as well there's going to be time for final thoughts and any questions that you might have. What is computational social science? It's the use of computational and empirical methods to address social science questions so we can break this down a little bit more and we can think about computational social science requiring some sort of human thinking and what I mean by this is you're going to need that social science brain that you use to identify important research questions and because we're dealing with social science questions we're going to need to understand all that social science stuff about how people behave and what they want and what they want to achieve so we're going to need that sort of thinking in order to formulate those really interesting research questions but as you can see we're also going to need a little sprinkle of computer thinking and that's so that we can turn these research questions into computational or empirical methods and after your research you're going to need that social science thinking again in order to effectively communicate those results to other people. Another way to understand computational social science is to think about what its main components are so let's go through a few points about what computational social science isn't so it's not just using computers within a social science research project you know I'm sure unfortunately all of us have had to send many emails in our lifetime but that doesn't necessarily mean that we've been doing computational social science and it's also not just using digital versions of purely traditional social science methods so you know it's not just asking for survey participants to complete an online form say on something like SurveyMonkey you know as opposed to like handing out a physical copy of one so yeah very important to stress that it's not just using a digital version of something that you would normally do in social science and it's not just using digital but purely non-empirical methods so I'm going to give you a few examples now that will untangle these ideas a bit more because you know I know they're not always entirely obvious so here are some examples that might help put this into context a bit so one example of a CSS project would be collecting processing and analyzing millions of online news articles to show change in political attitudes so you can see how we have that social science thinking here in order to formulate our research question about political attitudes but we also have that computer thinking inherent in using this computational method so in this case it would be web scraping and that's what we use to talk with the internet and gather a massive amount of news articles we've got a second example where we can use real-time weather and traffic data to explore how travelers react to events for example maybe you hear that there's a storm that's caused damage to a local town and you want to look at how people react and how that event is dealt with in real time so you can see how this project isn't just a case of using a computer with a social science project or using a digital version of traditional social science methods instead we have these uniquely computational methods and we have some other cool examples as well which include combining data from novel wearables or apps where you could establish a correlation between social media activity and heart rate so that would be aiming to answer a social science question about how people feel about certain images so you know you could look at how they react are these positive or negative feelings are they irritated and so on a final example is we could be interested in mapping family names over time by importing processing and formatting centuries of parish records that could allow you to explore the movement of certain families to different areas or you could maybe look at the changes in family size or whether families have moved away or not so these are just a few kinds of broad examples of css projects so moving on as well there are key factors that make a css project computational and the first one of these is data volume complexity speed difficulty or novelty and this is more important than the exact data source or type so in our previous example where we had parish records the source of data is not entirely important it's about the volume and the complexity and you know that difficulty or novelty of that data additionally another key factor like i've mentioned before is that the data must pertain to people actions behaviors choices and statements that's that key social science ingredient that we're going to need and just to hone in on that point the research question should be of course a social science research question which is going to use this atypical data to talk about how people make decisions or what influences their behavior and choices so the exact research question is not important but it must be within that social science remit and that's where we have this really obvious intersection between the computation and the social science and that's what we're going to focus on today so we have a nice little quote here um in essence css is an opportunity to do social socially valuable research that would not be possible without computational methods and tools so by this i mean that we couldn't for example manually scan years and years of police recorded statistics if we wanted for instance to try and understand how crime rates have changed over the last 10 years you know this physically wouldn't be possible to do manually and the same with you know centuries of parish records right um but with the use of computers we can apply advanced statistics and models in order to understand the change in crime rates or you know the how families have moved around different areas if we focus on that crime example we could look at how we could count many types of different types of crimes that have taken place in certain areas then we could aggregate these crimes to map them spatially we could look at exploring uh the long-term trends and seasonality or the noise components of these different crime types and this type of research you know it just wouldn't be possible without the use of those computational methods and another another example could be as i mentioned before web scraping millions of online articles and you could be doing that to try and understand how political opinions have changed over 20 years in order to get a sense of political opinions we'd want to also examine the words and the articles and how many words belong to different categories and what sort of themes are emerging and we could look at the proportions of these words and then how that changes over time so that's something that we won't be able to do without that combination of computational methods like web scraping and natural language processing okay so in order to see if any of that has sort of made sense we're going to have a little bit more interaction now what i'm going to do is i'm going to give you you guys the chance to vote on whether you think a given project is a computational social science project or not so if you want to head back over to menti meter we can get this set up and you can start voting so you can see that the code is on top of the screen um and also pop it back in the chat um yeah thanks emma um so yeah if you just head to menti pop that code in and we'll head to our first example our first example involves a project a project where we want to scan historic recipes and use ai algorithms to recognize text with the aim of identifying ingredients and measures used over time so what do we think is it css so we've got someone straight out the gate saying definitely css i'm gonna wait for a few more people to vote so yeah first some people are saying there's not enough social science actually a little a little bit longer interesting that no one said that there's not enough um sort of like computational um components here to make a css and i definitely agree we definitely have computational methods being used here but um i suppose that an issue could be that you know it doesn't seem to be a social science project you know um we're not there's not really a clear question as to whether we're asking anything about people's behaviors and choices you know it probably could be a css project but you know we don't really have enough information because it's not clear what the social science question is here so you know are we talking about how i don't know imported foreign foods integrate into domestic scenes and how people then take that open you know it leads to more wealthier people eating fancy foreign things and that trickles down and then those foods become normalized and accepted so you can see how there could be a social science question inherent there but you know perhaps we're just talking about the you know actual proportions of natural foods that people are eating over time which could be more of a biology question so yeah i know it's a bit of a wishy washy thing to say but there is no real right or wrong answer on this one this is just to help us explore different methods in css so i would i would probably say we might need a little bit more information on this one okay um what about this one um i wonder if you'll find this one a bit more clear cut or not so is this css or not so what we're doing here is we're using gamified smart home displays in order to understand how people interact with energy saving technologies so again right at the gate we've got quite a few people saying it's definitely css um personally i'd have to agree with that but again you know the question is it does depend it is hard to capture the nuances of research projects you know just one sentence so you know it's hard to know what the purpose is and what you expect the conclusions to show um we've definitely nailed that computational component so we've got this gamified smart home display that's probably going to capture a lot of digital data about how quickly people react in terms of you know what time of day it is or who it is that's interacting with it it could also possibly act as some sort of tracking data so that's going to make it you know fairly computational and if the purpose is to try and change people's behavior to make them engage more with energy saving technologies then that's definitely you know within that social science remit so this is where we circle back to how we conceptualize our thoughts and our processes but it's interesting to see that there's not as much division as compared to that last example and i'd say you know i've definitely leaned towards this being more css than the first example okay so what about this one um we've got a project that involves advertising for survey participation on social media with the responses then being stored in a database what do you guys think interesting so a lot of people saying that there's um not enough computation going on here very interesting and i'd also agree with that so if you remember on a previous slide we talked about how css is not just using digitized versions of traditional survey methods so you know um again a bit wishy washy to say there's no clear line between how much computation is enough computation but you've got to ask yourself if you can do the research um without digital tools then it's probably not css and if we draw the conversation you know back to surveys in itself typically the larger the survey sample the more necessity there'll be for computational tools because the analysis is going to become much harder but the use of online surveys it's not automatically going to make it a css project as you could probably you know if if you were to do it you know in person you could probably collect the same amount of responses for your paper surveys um and it's interesting as well that there's um quite a few people saying um they need more information to decide um because that's fair and that we don't know where the research is going from here and there's some people that said not enough social science very valid too because we don't know what this research topic is focusing on it could be you know just um collecting responses about participants weight and height for instance right so very fair enough answer let's go on our second to last one here so if you're feeling a bit questioned up don't worry we're going to be done with this soon so this example looks at reading in real-time weather and air pollution data to create complex models of hyper local air quality so let's see we've got people saying that few people saying it's definitely css majority of you guys are saying that there's not enough social science so yeah it looks like the majority agreed that this is not css due to that lack of social science focus um so we'd have to think about you know is there a link that's been explored here between hyper local air quality and human behavior and yeah I don't think and there isn't one implied as far as we could see so we would have to know more about the project are we looking at a political impact or is there going to be an overlap between different research fields and schools maybe with you know geoscience maybe or with physics it's not clear enough by the statement that there's an intent to explore human behavior so to say I'd agree there's there's probably not enough social science in this statement but yeah um interesting to see that there's not as much of a division compared to that first example but we do have a few people saying that there's not enough computation I'd say that the in my opinion I think that that probably is because we're looking at um reading in real-time weather and air pollution data to create those complex models at the air quality so maybe I'll think on that and we've got our last example so here what we're looking to do is train a neural network on social media data with the aim of creating a believable chatbot that can then counteract online radicalization so what do we think is this css or not I'll just wait a little minute for my answers to come through so yeah we've got a few people saying they need more information which is fair enough it looks like yeah nearly half you would say this is definitely css again I think I'd have to agree with you it seems to include all the components of the css project so we will just go through it a bit more and so it does seem that there is an implicit social science research question because we're focusing on people's actions attitudes and behaviors and we've got this goal here right that we want to counteract um online radicalization and we have a computational method to help us explore this so we've got a neural network which has been trained on social media data for those of you that might be unfamiliar with this term neural networks are a kind of artificial intelligence that teaches computers to process data in a way that's inspired by the human brain it's a type of machine learning process also called deep learning and it uses interconnected nodes or neurons and a layered structure that resembles the brain this is implemented in the hopes that we can use it to predict a user's behavior based on their social media posts and their messages so in order to train this model there's probably been a fair bit of web scraping involved so very computational stuff and in order to counter online radicalization we of course need to have good ideas about what online radicalization is and how people are radicalized and what kind of like mental processes are going on and I also think you know we've got to understand how people interact with each other because you know we're looking at how we can convince people to accept an idea or a question if we're going to make an effective chatbot so there's methods that need to be put in place to help understand how people change and this could be from using natural language processes so we could use this to get an idea about the breakdown of language that could be one example so yeah if we're using these kinds of complex computational methods to mimic and interact with individuals and in aims of you know sort of combating online radicalization then this to me is you know the quintessential CSS project that's our last example and thanks to everyone for taking part I think it's pretty useful just to explore the sort of components that do make up a CSS project so what we're going to do now is just go on to a little word cloud so go ahead if you want to enter up to three words about you know maybe what you've learned about CSS so far or maybe something that surprised you or something that you found interesting or anything that you're still curious about everyone will then be able to see your responses as they build up and sort of float through okay we've got machine learning behavior programming skills for sure um unclear confusing fair enough um hopefully it becomes a bit like confusing as we go on but I'll do my best big data yeah not just digital yeah what software is used computation that's coming up a lot advanced stats data scraping yep vague fair enough um natural language processing dialogue maybe I'll leave that a few more seconds see if anything else comes through function treat ethics scale yeah for sure research question is key yep need to have that social science element okay I'm going to leave it there um so we're going to talk a bit about how to become a computational social scientist and first we'll start by covering what a social scientist is so might be a lot to say but social scientists think like people and what I mean by this is that they use a lot of those like inherently you know human type thinking skills like abstraction inference a willingness to approach and understand fuzzy sort of concepts and background knowledge um not shying away from gray areas or overlapping categories these are just part and parcel of being a social scientist and it's no surprise that we need these sort of human type thinking skills because we're going to be studying people interactions and behaviors and this requires a certain skill set but social scientists um it's important to note that they do also build up a lot of data skills in the course of their research if you think about things like response categorization and coding quality evaluation pattern detection in statistics these are all things that you know many of us have done for our our um you know studies or our careers but whilst um you know a lot of us might have used computers this doesn't involve um writing computer code for most of us instead it might involve computer programs such as SPSS or STATA for statistical analysis but like I said um you know not much code whereas computer scientists um you know we are making some big big generalizations here but in contrast to the way that social scientists sort of think and approach their problems um we can say that computer scientists are taught more along the lines of thinking like computers and they need to think about how the computer is going to process the code and the thinking skills that they have are more along the lines of working with concrete definitions and absolutes so they're going to think in terms of strict hierarchies and categories in clearly defined scoped and scoped variables and rules and in terms of data skills computer scientists collect analyze and manipulate data through programming scripts computational methods and technological tools but unlike social scientists um again you know big generalization but they're not usually taught to identify or motivate research projects on the basis of societal impact so those of you that are social scientists might have had to justify your research on the grounds that it'll make the world a better place you know it could be I'm researching radicalization in online forums in order to produce insights that could lead to countermeasures whereas a computer scientist might be more focused on a sort of a more logical justification for a particular project um so it could be I want to make this algorithm more efficient so that it uses less memory so in order to do computational social science you're going to need a blend of those sort of two different ways of approaching a problem you're also going to need a great deal of open mindedness and you're going to need a mixed problem to work on so yeah you're going to need those skills that we've briefly touched on before and these are skills like being able to identify important problems or knowledge gaps considering possible solutions connecting problems to relevant theories or perspectives and being able to collect relevant information and research to frame your approach and these are all things that social scientists excel at so that ability to understand context and nuanced perspectives how to communicate abstract ideas and how to attack a research question whereas this is an area where computer scientists might struggle a bit as they're more used to um those like concrete definitions and clearly scripted variables and absolutes rather than those sort of gray areas or murky social science concepts so yeah that's going to be the first thing that you need um but you're also going to need um you know a bit of that computer thinking and what we mean by that is that you'll need the following skills so the ability to access organize process and handle vast or complex data how to write collaborative code um because the thing about um you know doing computational social science is that um you know you're not expected there's no reason why you have to code everything yourself right um a great part about this field is the collaboration that has occurred across different fields so learning how to write collaborative code as someone else can be really useful and how to properly document your workflow which is often a step that a lot of people do neglect so thinking about things like version control software and so there's one called github which is um used by a lot of people who write code and you can sort of upload your code with a little message each time and it keeps a history of all those different messages so you can see how the work is involved that's really useful so these are all skills that computer scientists might find a bit easier because you know they've been doing it for a lot longer as well as data scientists that have been trained in computational methods but for social scientists um it can be a bit harder to transition towards these computer skills but important to remember like I've said that a lot of social scientists do have those data skills that they can build upon and those are those things that we mentioned before such as you know coding responses pattern detection and statistics for mining surveys these are all quite technical things and you know learning a software like sbss or stata it's not easy it's at first when you you know you get to it but um you know it's it's much like um how it is when you first start learning the programming language and this is really important too um I can tell you someone that's moved from a social science field into doing more and CSS type stuff it can be a little bit intimidating at first um which is why it's good to remember that no one starts with all the skills that they need nor do they know all the skills that they might need to acquire and that does happen quite often so you know you might start off saying I want to scrape tweets for information on the 2016 us election and you expect well I'm going to need some coding ability for that but you don't know that that entails learning about different apis and file formats or you know ways to visualize the data but um yeah if you approach CSS with an open mind and a willingness to learn you're going to gain those skills along the way and you'll also start to understand that some skills have a steeper learning curve than others and that's going to lead you hopefully to collaborate with those from other fields and that's where we start to see this bridge emerging between the social science and the computing world the social scientists learn more about computing and vice versa we begin to see more conversation and collaboration between those two fields and what you'll find is you know you don't need to know everything about computer science but you're going to want to know enough that you can have those productive sort of conversations with people in those fields and finally you're going to need a problem that requires those CSS skills and that's a mixed problem so one that requires that that blend of social science and sort of compute science thinking so some of you might be here because you've already encountered one of these problems and I won't be surprised because as resources become more digitized these problems will become more important we've got the fact that you know interactions objects and processes have become smart or networked large volumes of data are now being made available and they're updated much faster than they used to be and there's so there's much more opportunity now for interesting research if you think about a classic social science problem so you know maybe we're interested in how men and women move through cities differently or how people with disabilities navigate cities so traditionally for this kind of social science problem you might station some interviewers in different places to stop people as they go past or to count how many people go by that are using mobility aids or you know maybe send out some surveys to people's houses but now there's much more opportunity for us to gather a large amount of data with computational methods so you could collect data from public transport networks about how many people actually bought tickets or how many people have swiped their car at a tram stop you can't even get you know sensors which can track how many cars or vehicles go past a certain point and you could use AI algorithms and CCTV cameras to identify how many people are moving for a space and even at what speed so as you can see there are new ways of approaching traditional social science questions but there's no reason you know why you have to abandon traditional methods completely and part of really the research is evaluating different methods and comparing outcomes and it might be interesting to see whether by using different methods you get different answers and if so you can then ask well why is that which then goes on to prompt further questions and we do different avenues to explore down so yeah it can be a difficult task but it has a load of benefits in terms of building upon your computational skills and strengthening those social science skills that you might already have so we are going to take a little short break just in a minute but I'd like to quickly introduce the eight step process that we're going to go through after the break and these steps involve identifying problems exploring the problems formalizing concepts collecting data implementing software and verifying the concepts then using these concepts to experiment or analyze data discussing your findings and presenting a conclusion then communicating publishing and presenting your findings and sharing them to wider audiences as well as documenting and validating your findings so to make this process useful to you what I want you to do is to start thinking about either a project that you'd like to tackle or a research idea that you've been thinking about or it can even be a project that you've done in the past so you can jot this idea down or maybe even put it in the chat and just have it in mind as we go through these steps so once you've identified the problem you're going to want to be as clear and specific as possible about the pattern the problem or the lack of insight you'll also want to identify who is involved where it is etc and what this will do is it'll help you define your research question so maybe we have a goal in mind of we want to get more people traveling actively through city centres we want less cars on the road and more people riding their bikes or their scooters or you know just being able to get from A to B in their wheelchair the research question might be what are the barriers to active travel in city centres so what you then need to do is you know identify who's involved so you can start to list who is involved whether that be potential companies people or different demographics that might be of interest to you so you might want to look at city councils, bus companies, different businesses, vehicles it's better to just you know as well just go all out with these lists too as it's going to give you lots of different avenues to explore and you can always cross out any that you once you've done a bit more investigation into them or decided that they're not actually that relevant so yeah first thing first identify the problem the next step is going to involve exploring the problem and this is where you'll gather information and perspectives in multiple ways so surveys, observations, secondary data analysis, web scraping, APIs so this might involve conducting a few interviews with people of interest you could interview the manager of your city's transport network or local council workers but you would probably also need a survey or observations or secondary data analysis to capture how many people are actually moving through the city centre so it's about using different methods and tools to further enhance your understanding of the problem or the research topic you'll also want to spell out sub-problems, processes, relationships, simplifications, assumptions, related issues etc so after settling on your main research question you're going to need to get more specific in order to make that question relevant and measurable so if we take that question what are the barriers to active travel in the city centre let's say that's our main question you might specifically be focusing on you know what are the barriers to active travel through this specific city centre at this time of day given the way that the roads are laid out so yeah this is where you really nail down the particulars of your research question so we'll just switch over to MentiMeter for a minute where you can feel great to tell us a bit about your step one and two and you don't have to share if you don't want to and you can always just keep a note of it for yourself but you might just want to get some of your own research ideas floating and just to note as well it doesn't have to be you know super well defined I know it's hard to outline these these steps if you've only got sort of a vague idea about what you want to do but for step one it could be a simple as people on Twitter from different political camps really seem to hate each other you know so I want to look at political polarization on Twitter or X and your step two might be I'm going to gain access to a Twitter's API and see what I can find by scraping some tweets and then maybe I'll see how I can build a big data set because anyone have any step one and two that they want to share don't worry if not we can just move on or you can put it in the chart what are the social determinants of mental health okay good yeah so we've got you've identified the problem and any ideas about what sort of computational methods you might look at to explore that and you can pop them in influence our media attention on report in Romance Broad and we've got a step one here so what are some of the practical barriers for people accessing mental health support step two interview interview staff surveys using appointment data nice so we've got any ideas about our sort of computational methods that we might want to use and ways that we might do that and beyond that we've got comparison of policy making processes across UK to bold jurisdictions very interesting a good question here um our chat box making users more depressed so we can think about some natural language processing there um sort of looking for the proportion of certain terms over time um maybe some sentiment analysis as well so sentiment analysis is quite cool you can take a sentence and you can basically get a polarity score um which will you know sort of measure sort of measure um how negative it is or how positive it is what are the elements in the language used in social media which may support uh racist discourse nice so again if you're looking at language and natural language processing is pretty good text mining tech techniques so you could look at maybe there's a lot of apis for different social media um forums and stuff I know x or twitter has unfortunately now has gate kept its free api too it's now behind you know a bit of a paywall if you actually want to scrape people's tweets which is you know a bit of a shame um but you know although I think there's definitely um apis for stuff like reddit which um you know could be a good source of um you know getting info about um social media which might might support racist discourse or maybe you want to look at like um some of the more I guess darker sites like boar chan or other um message boards that can be a bit controversial and sort of like have um those like um more radical sort of cultures um determining the tendency of judges following advice opinions submitted by their support team scraping data off published pdfs using um machine learning to extract info and sentiment yeah that's a really really cool idea and there's um actually a github repo on our UK data service open account that um my colleague um Jules has made public and she's um scraped a lot of info from pdfs but I know that she had a really hard time um scraping pdfs so um if you are interested in that I can put it in the chat um but yeah we've got a bunch of stuff on our github as well just to put that out there that um we've got stuff on text mining natural language processing machine learning this is some stuff on clustering but yeah um scraping data off published pdfs using machine learning to extract info and sentiment good stuff um how an entrepreneurial's narcissism can influence funding success in crowdfunding platforms nice so again um you know maybe a bit of web scraping for that um nice so yeah we've got some pretty good stuff um thanks for um all the answers um we'll now go on to our step three so so now we've got step three um and this is where we come to formalizing um our concepts and what I mean by this is you're gonna want to make the concepts and processes explicit formal and both computer and human understandable so you know computer scientists often refer to this as pseudo code where we sort of write out our um recipe for how we want um our code to go um but you don't need to know how to write code you just need to start understanding how to formalize things so you know when we talk about social science and those like murky and social science concepts um we could look at something like a research question which focuses on thrust okay so that is like I said very social science concept and if your goal is to get a computer to be able to measure it or model trust or represent it in a simulation you're going to have to define it in a way that a computer would understand so maybe you'll define trust as a variable between zero and 100 and you're going to want to make rules then about how that variable will change in certain situations so maybe if two parties interact positively thrust increases and then given a negative interaction um level of trust will decline or reset to zero so you know you have to start thinking about how how we formalize concepts in our research questions so that a computer would be able to understand it so you know something like um we talked about sentiment analysis right where we give a score to certain words in a sentence or a text um and we give them a score from you know zero to one based on how positive or negative that sentiment is so it's about thinking about how we sort of translate those concepts into something that um a computer would understand and then you know we have that computational element so I hope that um that bit of formalizing our concepts makes sense so um now we're on to step four which involves collecting data implementing software and verifying your processes so what you'll need to do is select and implement one or more methods so many of you might have thought about these methods in step two so we saw a bit of that before some of you might have wrote down web scraping and maybe you're thinking about sentiment analysis or agent based modeling there's a ton of different you know when it comes to computational methods there's a ton but this is the step where you're going to implement these methods and make sure that they work in the way you anticipated it so um you know when you're working with um sort of like um computational data you know all you've heard and you're messing around with it um it's um it's never always as you expect right so maybe the data comes in a different format than you expected for you know like I said before with someone mentioned a pediatric you think well um probably not that much different than scraping a web page or a word doc no it's for some reason much more um complicated and it involves um you know learning a little bit more about a different approach to doing things so you know there's a lot that can be unexpected or maybe you know you're encountering a lot of error messages in your code so this sort of stuff is about you know experimenting and um making sure that you can make it work in the way that you've expected so you could instead of using your full data set while you find your method you could use a toy data set or a reduced one I often do that if I'm having trouble sort of um with a function that um I'm working with um it helps me sort of like better process things if I reduced on my data set to like a few rows and I can think okay what's it doing on each row so it's all about fine tuning on methods and of course as well the choice of method is going to be highly dependent on the research topic too and lastly you're going to need to thoroughly check um the selective method has been implemented correctly so this is what we mean when we talk about verifying your process or method it's about answering that question did we do the thing right okay so moving on to step five you're going to want to experiment and analyze the data so of course you're going to run the experiment build any of the models that you need analyze the data or otherwise use the methods selected in the previous step and you'll need to identify and explain the results within the context of the experiments and the model or the method that you're using then once you've run your experiments and you've analyzed the data you can then start to interrogate the results and form some conclusions so this means going beyond the experiment and the model or the method that you've used to draw some conclusions about what the results mean so what sort of picture are your results forming do you'll find and support any policy recommendations and um so you know is the research indicating that there might be certain suggestions that you can make maybe to local government or something and you know then we have who or what these results affect and why does it matter and what should change and who benefits from those proposed changes and so if we use that example of looking at how people move through cities and perhaps you found that people with disabilities relate to mobility have more difficulty navigating through particular areas maybe there's uneven surfaces or particular rows that are too narrow in which case then you know the recommended changes would maybe include wider footpaths in x y z area right and we are near the end of the research process now where we're going to be focusing on communicating and sharing that research and in terms of communicating the research it's important to understand that all of the previous steps must be communicated to multiple audiences in multiple ways and what's really important is that you think about short-term and long-term engagement for a lot of us our mind will instantly go to you know well i want to get it published in an academic journal which you know for sure definitely important but it's also good to think about other forms of communication so are you going to present your work at a particular conference or submit a piece about it to a blog and you can think then about whether there's any sort of workshops or classes that you could present on it and sort of get your research shared more widely and with that you know as well that comes with thinking about how you adapt your style and your tone of communication to suit those different audiences for instance you know presenting at a conference is going to be much in a much different style and tone maybe then submitting it to a more sort of casual blog especially when you're talking about you know computational stuff you're going to want to think about how you communicate that quite clearly to someone who's maybe thinking well you know what the heck is social network analysis like what what do these nodes mean and what are these connections showing so yeah think about how you want to communicate that research to different audiences so yeah um finally and perhaps most importantly you're going to want to share document and validate your findings and what we mean when we talk about validating your research is making sure that the right thing was done by allowing your work to be studied reproduced and or modified as needed so I'll just go back to that chat message said that you know increasingly there's a need to demonstrate how we conduct the research how does documenting workflow operate now when you're working with computational methods a lot of the time you're going to be implementing these methods in code and so one of the sort of like big hurdles of first getting to computational sort of stuff is getting your head around github now github is what's known as version control software so it logs a copy of your code each time you submit a change on it and you sort of like push that change out with a little message so you can then see it'll give you like two side by side versions of your code and a new bit so stuff that you've added and what the code looked like before so you can track it as it evolves and a good documentation as well when your coding includes intelligible and concise code comments so when we're coding we can use a little hashtag symbol and it allows you to add a line of text that doesn't get executed when we run the code so you can add you know you can say like head function shows the first five rows of data you know you can use these you know code to describe what's going on what functions you're applying so github is a really you know a good thing to get your head around you it might involve learning a little bit of stuff about the command line too so and you know how you interact with your operating system and where you are in the computer I know it can sound massively confusing but github is something that I would definitely recommend it's great for uploading your code and letting people see how it evolved so yeah you're gonna want to allow as many people as possible to be able to access your methodology in any code or data and so you know you might look at getting a zinodo doi so digital object identifier for your github as well and you can just keep everything there and you know your article that you've done and any sort of links and you know you're gonna want your research to be as transparent and as well documented as possible and of course there will be caveats to this so some of you might want to work with admin data that's restricted so you could look at work around so if you are working with admin data you could maybe create a dummy data set so that people can still run your code and work through your methodology but without you compromising any sort of sensitive data and yeah just to note on that the reproducibility is a really important area of research I don't know whether some of you might have heard of the reproducibility crisis or the replication crisis that's going on in the social sciences and psychology and stuff and there's been a big push for the social sciences to sort of like like was it like Suzanne said to you know sort of increasingly demonstrate how you know you sort of your methodology and stuff so it is really important and it is unfortunately often neglected which is why it's good to think about how you'll document your work before your research gets underway and there's just a few things to note as we've had towards the end of the presentation and that's the steps that I've outlined so these eight steps they're not linear you know there's going to be many points from each step that you'll need to return to or apply throughout the research process so for instance documenting your research is something that you'll want to apply from the beginning and you know it's annoying when you sort of go back and you realize oh I haven't added code comments or you know I haven't got a lot of documentation for this and you have to sort of like go back through your code and think okay well what did I create this function for what's this doing you know it's just so much easier to add code comments and documentation at the start and when it comes to computational social science projects most or all of them are going to require many iterations right and that means you're going to need to revisit certain steps so maybe you'll come up with a research question in step one but after you've actually explored that problem further in step two you might want to then jump back to step one to then reformulate that research question in light of something that you've read or maybe you're on step four right and you know and you're you've implemented your method but then you need to go back to step three because you've not outlined the concepts and the processes so it's a very like iterative very iterative process and yeah like I said I've often had it where I'll make some really important changes to my code I don't document it and then when I come to write my methodology I'm then struggling to explain why I opted for this type of algorithm or this particular Python package because you know my documentation isn't there and that's a shame then because what happens is we lose those really interesting insights right so you know it helps us figure out why someone took a certain path and you know why did they use scikit learn instead of I don't remember the other machine learning package but you know just document you got everything there then that you need so before we head to the end I'm just going to go back to the Mentimeter because we've got you know just one more I think it's like a word cloud but yeah just to note that this is pretty much the end of the presentation and then we'll have our Q&A but I'll just head to Mentimeter where we'll do these last interactions so we've only I was going to say we've only got a few minutes left but we're actually ahead of schedule so if you want to share your last minute takeaways on what CSS is and why it's important and you can pop them into Mentimeter what if you don't know how to code or even read code and yeah that's absolutely fair enough it is a big hurdle and you know learning computational methods and unfortunately a lot of the time does involve having to learn a bit of and sort of like simple programming skills whether that's in the two most common code languages use are or Python and that I've come across in this field and if you don't know how to code or even read the code fair enough and I suggest that's an area to start with and I will link our GitHub so what we have on our GitHub is we have these sort of they're called binder files and they allow you basically if you don't have any sort of like programming languages installed or any coding editors installed they allow you to run the code just very simply just by pressing a sort of like run button and you can go through our coding scripts that we've designed to be like very simple and like user friendly not nothing crazy so that would be a good place to start and I am also going to put that I'll put our GitHub in general and then I'm also going to put the GitHub repo for the person that was looking to scrape PDFs but yeah let's just search about what people have said questionable fair enough learn new skills document consistently need more time to digest fair enough need collab networking yeah I will say that's why the CSS problems are very good because we have people that come from all sorts of different skill levels and you know we have people that are just they don't know where to start a bit like I was back in the day you can come along to our drop-in session we can you know help you out give you like you know one-on-one sort of like help that you might need and so yeah next one of those is folding it with me at one o'clock you can sign up on our UK DS events page and we have loads of people that come to a super like willing as well to like offer their advice as well so it's always like a really very nice support of atmosphere and confusing yep fair enough need good info sources iteration yet for sure getting many different iterations of those eight steps that very important to remember like because you're iterating over them they're not linear and you're going to have to go back and refine stuff document consistently yep glad that I nailed in that and multi-functioning yet excited to learn more cool yeah I'm really about to glad to hear that and usual usual research usual research process yeah I mean fair enough because hopefully I wanted to put like a CSS sort of slang on it but yeah it is obviously you're following like a normal sort of research process but I wanted to sort of illustrate how that would go if you were doing something CSS no the limits yet support needed definitely definitely understand that I've been that person who's you know looking to start getting into coding and has not known where to start and that's why it's going to follow through and you know github repose and stuff like that and we will get to the Q&A I'm going to move on just to before we end as well there's a few references for anyone that's interested it's my colleague Jules that designed this talk so just shout out her and just say as well that we'll be making these presentation slides available so you can always go back and look at these in more detail and it's all going to be uploaded online anyway so you're going to have that recording if you need it and if you need to go back and check on you can do and I'm also going to put up my contact details as well so I've got a X account there's also my email there too so if you want to ask me anything about the workshop and or anything more CSS related and you know just get in touch and like you said as well I know I've said it many times now but we do have those CSS drop-ins and they're really useful if you want to connect with other people in different fields that are doing that sort of stuff they're really useful and the discussions are always interesting thanks so much everyone this has been great