 So this is a session, really, I suppose this is filling a gap between our expectations of when sensors 2021 will be released and when it's actually going to be released. So the latest information we have is that this is going to be available from October and I'll go through that in the session in a bit more detail. Please feel free to raise questions as we go along. There's quite a lot of information in here. And if you have particular interest, I'm happy to stop at different points. Ideally using the Q&A section so that we can pick them out easily. So what I'm going to cover today is what we know so far of ONS release plans. I think we expect these to be firmed up with some more dates fairly soon, but there's nothing when I checked yesterday to update what we have. Then to talk about some considerations about using sensors data and finally to share with you what we have to offer in terms of events like this and other events that are planned to support your use of the sensors. First of all, there's a phase release over a year. So I'm going to go through each of these in more detail, but phase one, we're kind of in at the moment. So on 28th of June, we received the local authority summaries by age and sex from October to December, we're going to get the topic summaries and area profiles. They cover broadly the same areas, but the area profiles are more interactive, whereas the topic summaries are intended to be more downloadable, I think, if you want to think about the distinction between them. Phase two will include multivariate data. So that will be a set of tables predefined as in the 2011 release, though there are less tables in this release currently in the schedule that ONS have published. And alongside that a flexible table builder, which will be available via ONS. So that will enable you to populate tables. We'll say a little bit more about that later on. And they've also brought forward one of the short term, the small population data sets, which is for short term measurements. Phase three will be a set of other data. So more different population bases, more small populations and flow data and micro data. This is really based on work with ONS, so it links to England and Wales. Northern Ireland actually released the first data and they've got another release planned this week, but there's a slower schedule for the release of their data. And as you probably know, Scotland, the Scotland census only completed a year later is extended by a month. And so we don't expect results from that until well into 2023. Once all of the results are in from England, Wales, Northern Ireland, Scotland, there will be a UK wide census release. But there will be if anybody's interested in that, ask me a question about it later on, and we can talk about what will be included and some of the challenges in looking at that data. So the topic summaries are going to be mainly univariate data that will be released every two to three weeks from October onwards. So the first block are on demography and migration. They include variables like the age and sex breakdowns, but at different geographical scales. So down to output area level, administrative scales, etc. I'll talk a bit more about the geography in a minute. They will also include country of birth, age of arrival and year of arrival in the UK. So we'll cover migration topics as well. The second group will cover ethnic group, national identity, language and religion. The third, the veterans information, which has been brought forward. The fourth on health disability and on paid care. The fifth on housing, which includes information at household level about tenure about overcrowding, etc. Then labor market and travel to work. The next area, these are voluntary questions. So they're new to census 2021, new to the census altogether. We're not clear on the return rates at the moment, but they cover sexual orientation and gender identity. And then finally, education. So this data is available for there will be variables for individuals, variables for households, and also in terms of housing variables for dwelling spaces that will show something, some estimate of empty dwellings. So Suzanne's asked about returns. Initially, in Scotland, the return rate was quite low. I think they've largely increased that by the extension of the month. So there will be some issues, I think, particularly in Glasgow that they expect, but I think they're reasonably competent. They can make population estimates based on the returns that they've received. I think we're waiting for more information from Scotland about how that process is going now. So they have actually changed the way that they're using administrative data to check the returns. Right, so in terms of geographical data, this is now available from the open NS Open geography portal and will be available from the UK data service. It has been scheduled to be loaded up. We still have some work to do on our website to make 2021 census data visible and that will be done by the time the data is released. The full sets of area and bivariate data, multivariate data. So the boundary data that's there is administrative data on local authority health districts, electoral data on areas like constituency awards. I'll come to you in a minute, Robert, and statistical areas built up from output areas, which I'll say a bit more about afterwards. Robert's asked that we won't be able to tabulate health data against ethnicity. That's included in the multivariate data release. So my understanding is that for ethnicity, for example, what you will be able to see is ethnicity by age and sex, but not by any of the other variables until the multivariate data is released. I'm pretty certain that's one of the tables that is there. But if not, it should be available through the flexible table burner. Okay, so that's the geography, just to say a little bit more about output areas. So output areas were introduced in 2001. The idea was of them was to give a more stable geography than the kind of ward data had done previously. So significant reorganizations forced by changing populations meant the war boundaries have not really been consistent in many areas over time. So in 2001, output areas were built and they had a kind of set size with a minimum and maximum and a target size. They were based on an analysis of matching characteristics within the population so that the areas were more homogenous and subsequent revisions have tried to minimize changes. So my understanding is that the, well, the number of changes in from 2001 to 2011 was less than 5% and similarly that target has been met from the 2021 census. Okay, Cedric, yeah. So the univariate data, phase one data will be available down to output area level with one caveat about statistical disclosure, which I'll talk about later on. So those output area geographies are then built up into two and what's called a lower level or LSOA again with target sizes, and they are linked to quite a lot of data that's published. So, for example, the index of multiple deprivation and the different indicators, administrative data, and recorded crime data. And the next level up is mid-level super output areas, probably akin to the size of wards. Again, there are statistics published at this level. So GCSE attainment, education attainment, COVID cases. And all of those output area geography boundaries are constrained within local authorities. So we don't get overlapping areas between local authorities. So they're usable in that sense. Okay, so area profiles are similar and there's some examples in here of the types of things that are there. The first release on demography and migration. As I said before, there's sex and age. There's also legal partnership status, whether people are living in private housing or communal establishments, their country of birth, how long they've lived in the UK, and things about household composition. I think I've covered most of the others in terms of work. There's economic activity, hours worked in banded areas, industry and occupation, and also information about travel to work. Again, there's a caveat on that given that the census was taken mid-pandemic. There probably wasn't a lot of travel to work for many occupations. So what we have is limited to some extent. In terms of education, we've got whether people are in full-time education or highest qualification. And on health, there's a self-reported subjective health measure and also long-term health problem or disability, whether the person provided on paid care. So glance phase one. As I said, if you've got any questions, then feel free to put them into the chat now into the Q&A and I'll pick them up as we go along. So in phase two, we're going to get information about the short-term population. This is people who said they intended to stay in the UK less than 12 months. There will also be defined sets of tables, and we will share this version of the slides, but there are previous ones that give you that link to the details of those defined sets of tables. And they will be available at different geographical scales. In terms of the flexible table builder, all we've seen is a mock-up, but it's worth kind of diverting a little bit to explain what will happen. So in 2001, we saw this strange changing of numbers. So we didn't have one in two counts. We had zero or three, and that could mean any of those numbers at different geographical scales. In 2011, the challenges with that mock summing to the total were recognised, and what happened is that small counts were swapped between different areas. So small counts may not reflect the actual area they were in. This time that process will happen again, but once the table has been defined, there will be a rule-based statistical disclosure control that will suppress and tell you how much has been suppressed of the information you've requested. And I think that the last thing to say about that is the variable categories will vary between univariate and multivariate tables based on that. Now, at the moment, my understanding is that for some variables, and I'll use ethnicity as an example, a reduced set of categories will be available in the flexible table builder and in a lot of the multivariate table. So previously what happened with this data was that the balance between scale and the number of categories was determined by ONS and the pre-published tables were therefore limited in scope in particular ways. At the moment, it looks like the scope on variables like ethnicity will be caught back in the multivariate tables. I know there's been quite a bit of lobbying of ONS to get them to change this, so it's a bit of space we'll have to wait and see. But overall, I think the message is that you need to balance the geographical scale you're looking at against the level of detail you want in your categories. Okay, so a question from Samrie all about the multivariate tables. They will be down to output area level subject to statistical disclosure control. So depending on what you're looking at, if the counts are reasonable, they will be there. And there will be population and household estimates, I think is the thing to say. So in the data published in June, what we saw were the actual counts of data collected and also the population estimates derived from those alongside mid-year population estimates. I don't think we'll get that level of detail down to output area level, but there will be estimates of population and household at that level. Marion, the population estimates are the population estimates. That's what we will get, and that will be the base of all statistics released from the census. For those from local authorities, there's a strong commitment from ONS to look very carefully at mid-year population estimates for 2022. So though these won't have characteristics, they will be used to inform resource allocation because as you're probably aware, there's significant concerns about population change during the period of the pandemic. And some authorities very concerned about losing resources as a result. Okay, so, and then phase three, I'm going to go through each of these in turn. So these five areas, I haven't got much on detail migration data. The promise is that will be a number of tables developed around migration. And that's included there, but at the moment it's only words. There's no real, real details. So migration data broken down by a number of different characteristics. So the alternative population basis covers workplace population, workday population, which is a combination of those people who work in a particular geographical area, and those who aren't working and who live there. And out of term population. So this affects students. So students were asked to complete their return as if they were at their term time address, even if they were staying at home. And quite a lot of checking has been done to make sure to allocate students to the correct address. So the population should reflect students at their institutions and the address they would have been living at. If the pandemic hadn't happened. And it's the basis for kind of further analysis of our issues with students. The second address is where people have identified that they live have a second address which they stay at more than 30 days in the year. So things like holiday homes, but also work based addresses for people who have a residential and workplace living around term. In terms of small populations, we have a number of writing categories, and there's a there's a kind of promise, depending on statistical disclosure control to include a breakdown by five year age bands by sex for selected ethnic groups and countries of birth. So the categories that eligible for this are ethnic group. So I would sound a note of caution with ethnic group in that not everybody will have completed this for those who completed a tick box answer. There's no requirement to fill in ethnic group. In England and Wales in Scotland, interestingly, there were different ways of asking this question. All of them had another box right in box. But for example, the prompt for black African identified the country that people came from the the prompt for black Caribbean prompted a question about identity. So they will be there. They were very rich data down to our area level in 2011 and 2001, whether this statistical disclosure control will reduce the availability of that information is not clear at the moment. Similarly with country of birth, there's a lot of detail on country of birth. And in the raw variables, there's quite a lot of information. So we may well get a good look at that in phase one. Religion has a writing and national identity has a writing. There are a number of groups that based on people's requests, there will be data sets on MSO a level. So that includes Cornish Jane Kashmiri, Nepalese, Revadassia and seek. So the question about British sign language, and there will also be consideration about writing data and how robust is on other groups so they've identified Romanian Somalian and Turkish Cypriot. So flow data is probably different, but it is based on the origin origin of the flow and the destination of the flow. So it will cover issues on migration. So internal migration within England and Wales to start with that will be harmonized with Northern Ireland and Scotland once that data is available. Workplace flow so the commuting information that's previously used for calculating travel to work areas, etc. Second address flows. So for those people who have a second address where it is and their flows and stupid flow. So this is will be based on the term time and out of term address. Finally, in this section, the micro data samples. So these are data sets, the whole multiple variables for individuals. So they are a 5% sample regional and combined local authority geography. So the combined local authorities are relatively smallish number. So they are smaller, this small, smaller districts, the city of London, the art of selling as examples. So this will be for the first time in quite a while, a 1% sample of households, and there'll be a further 1% sample that's sent into an international database. So those are all safeguarded. I should have said the flow data and micro data are likely to be only available from the UK data service. On the flow data, there will be open data, safeguarded data and secure data. So there are three levels, open data, you don't need to be registered with us, you can just pick it up and use it. Examples include things like teaching data sets, but there isn't a lot of information in them. They tend to be quite limited relatively. Safeguarded data, you need to be registered with us and generally you can access safeguarded data. You do have requirements to notify the data owner of the purpose you're going to use that data for. And then secure data is the most detailed and you need to go through an accreditation process to become authenticated to use that data and then to use it in secure settings, which essentially means you can interact with the data but you can't take things out without them being approved. So it's quite resource intensive, both ONS and UK data service operate secure data training accreditation and also provide facilities for researchers to access that data. And in the secure data, there'll be a 10% sample of individuals and a 10% sample of households. So it's worth saying as an insurance policy for some of us, this might be a useful way and getting at more detail, multivariate data. Though the geography is more limited, we won't be able to go down to local neighborhoods within it, at least in the safeguarded data. So that's the kind of summary of what to expect from ONS and when what I wanted to spend a little bit of time thinking about is the considerations you need to take into account. So first of all, the obvious one, the pandemic was carried out during, sorry, the census was carried out during a pandemic. So there will be impacts on employment, things like employment, where people live, what work they were doing and education. And basically ONS are going to report fairly soon on an overall, if they haven't done already. And they will be looking at this in more detail. And they will provide supporting information alongside the releases of data saying how it might have affected particular variables. I'm sorry, I don't know the answer to this. I think there are some challenges with the rural urban indicator, but there is a, there is a kind of classification there. I suppose I can't answer that question. I'm sure it's there on the ONS website. And if you want to log a call with our help desk, which I'll give you details about later on, then somebody will pick it up and have a look at it. I can't do that interactively at the moment. At the time there was a new quality assurance process, which meant that the figures that were from the counts and the estimates that were being made were shared with local authorities who wanted to I think around 2017 England and Wales did so. That's one of the reasons for the delay because that process obviously took quite a long time. But what ONS asked for was evidence that would challenge the counts and estimates that have been collected from those local authorities. And there's already a kind of some information in draft and there will be publications from ONS on that process and how that worked. So I went to a conference recently where ONS were talking about these and there's a number of people working on publications to support our analysis in the future. So I think an important thing for us to think about is the coverage. So if we look at the coverage overall, the headline figure is very promising 90% percent. But that varies significantly between areas. And I suppose the areas we would expect to have lower rates of coverage, it may be the areas some of some more interested in. So the way that those counts have been turned into estimates might be important depending on the type of analysis you're doing. So you would need to think about and read the publications that come out around that process. Again, these are being worked on and will be released as we go along. But this kind of background technical information is quite useful because we are dealing with population household estimates. Then there is the changes in geography. So if you're interested in change over time, you may need to think about how the geography in the areas you're looking at has changed specifically. I did work on comparison between 2001 and 2011 census looking at Greater Manchester at neighborhood level. And there were significant changes caused by housing market changes. So the development of city centre and town centre living was quite significant. It's likely to be significant again in quite a lot of places. And also the clearance from housing, the changing nature of housing provision. This might be more challenging to get at because I think for me in 2001, 2011 what I saw were major redevelopments taking place. It feels like in the last 10 years we've had a lot more infill change. So whether that will impact boundaries or whether the boundaries will stay the same. I think that's quite interesting, but places will will inevitably grow in parts of the cultivation I might look at. So this is a question for Maria on English date English language data. So this is included within the second release at this to the language national identity and religion and has components around English language competence. Well, she also has things about Welsh language competence for Scotland. He has Gaelic Scots and for Northern Ireland, it has those Scots and Irish questions. So depending on which country you're looking at, they will all include something about English language. And it is the second release in the England and Wales census data so should be available this autumn. So the last point to make here is around trust and willingness to disclose some characteristics. So we are in a strange time. There is lots of lack of trust and particular groups may particularly distrust aspects of the questioning. So for some other work I was doing I was just reading on a report by government published this year on gypsies and travellers using 2011 census data. So they go on the baseline of the population then which according to the census was 55,000 around. In fact, what was the actual population based on a number of other sources like accommodation needs assessment was estimated to be 300,000. So a number of people from that community didn't feel willing to disclose that aspect of their identity. With voluntary questions we really don't know so on religion on sexual identity and gender identity, sexual orientation and gender identity. We don't know whether people would disclose things and we don't know what people might feel sensitive about so we've always held to not asking about citizenship in the census. But there may well be some lack of trust from those going through asylum processes or have been through them to disclose those characteristics. So I'm not sure if there are sensitive sensitivity issues about military veteran status. Yeah, I don't know how people feel so I think for those researching in the area, it's something they would need to consider but it's not an area I'm familiar with I'm more familiar with issues around race migration and that kind of set of data. Okay, so that's the kind of issues to consider. So the last part of the presentation is really thinking about what we have to offer. We're clearly not the only provider of training. It's worth noting that many of you probably already know this that we are funded by the ESRC. So the training we provide is free and the materials we provide on the website, etc, free to use. And we will be a key partner in we are a key partner in the delivery of census data and the repository for the more advanced products like flow data and micro data. So we know that some of these may be also be provided by ONS and NOMIS. So these are likely to be available from other sources aggregate data. So to be able to select data, most of you will be familiar with NOMIS or the UK data service. Our interface is being refined. So it will probably look more like the NOMIS type interface and that you will be able to select the variable and see what's available in a in a more friendly window interface than infuse infuse work fine. The geographical data will be available from us and from the ONS geography portal. And then, as I said, for flow data and micro data are likely only available from us in terms of thinking about what we do. I think we think about different audiences and are developing ideas about how we can promote access by a range of audiences. So our core is probably the academic research and teaching establishment. Many of you will come from that background. We are also working with students in HE, but trying to extend our reach out into schools, FE and schools and we will be providing census information and guidance for aimed at school students. In terms of the public sector, we have used by national departments, local health and councils, and we are thinking about gearing some products. I think we'll promote the data service interface fairly soon that the web designs changes are in and they will be pointing you to different places on the UK data service. Where the data is held at the moment they reflect kind of a lot of information that is not about just getting at the data that you want to use. So that redesign of the website is imminent and will provide access. So I'm not sure it's going to be that link you say specifically, but there will be front page access to census data. There will be information where you find census data that will reflect the priority that a lot of people are placing on 2021 data. So I said there's public sector interest and voluntary and community sector interest and we have been talking to key stakeholders in those sectors and thinking about how we can better reach them with materials we have, given that they're free to access. So obviously they have a cost to using them in terms of the organizational capacity. So the kind of things we're doing things like this so just telling people about it. I've done. I've had a series of presentations at conferences in the latest round, including one run by the body be trust which was quite interesting on race and migration. So connected with a non traditional audience for us. We're running a series of how to set sessions linked to the release of the data. So there is a geography session coming up. There's an aggregate data session plan so data and spotlight sensors. So we've talked at the end of October at the moment as flow data and micro data release next year. We'll also be looking at those and alongside that we're interested in thinking about techniques and substantive areas of interest. So, thinking about the kind of things that people might want to ask questions about, and starting from a problem rather than starting from an interface and access to data. So plan some sessions next year, jointly with our NS in some cases, where we're going to look at statistical disclosure control, the courage and imputation and the implications of COVID-19. So those are scheduled. Well, they're not, we haven't got dates yet, but we plan to hold those in spring. Sorry, from early 2023 onwards and the modifiable area aerial unit problem, which, for those of you who aren't aware is the impact of drawing boundaries around particular places and what happens on the edges of those boundaries. An interesting example in London is that the change towards means that the data is based from ONS on a particular way of estimating the GLA have refined that method to use a housing waiting based approach to allocating their ward counts. And in terms of training and materials, we have one more workshop on planning and research design schedule for next week, which is aiming at picking up people who have a clear research idea and want to take part in a kind of focus of thinking about how they might plan that we have a session in November on using census data for teaching and we've been working on developing Excel tools to demonstrate how you can analyze census data and also provide information on how you can create your own analysis using Excel. So I think for us, the branch into Excel is a new one relatively, but is as it's so widely used, particularly outside of the academy, it feels like a space we need to fill because that's the way many, many people will be accessing that data. We'll also be producing documents that will update the release schedule as we get information, doing short video materials to support the training, and we're working on a set of explainers for key census variables. So picking up the new variables and those are quite complex to explain what's there to give details of the categories available, how they're available, how they're planned to be included in multivariate analysis. And in some substantive areas, some of the cautions about using them for particular purposes. And finally, we have a help desk which you can email any query to and they are picked up by people like me, who will address questions you raise. The help desk is has broader functions so if you have issues about accessing the UK data service, etc, there are other parts of the service that will pick that up. But our team pick up the substantive queries will pick up the substantive queries on census issue.