 So, my name is Trisha and today I'll be speaking about building digital democracy solutions in Kenya through data, and why did I choose such a long title? But most importantly, I'll be talking about how these solutions contribute to governance accountability decision making for everyday citizens. Sorry for those who heard me earlier, but I have to introduce myself again. I'm Trisha, I manage Code for Africa's data lab, and I'm from a city called Durban, which is on the east coast of South Africa, which is relatively warm, so when I got here I was like freezing. Yeah, so the open data world found me, and we can talk about that later, what I mean. But what I love is that how open data and tech can be used to drive change and impact in Africa where I work, and the global south, now being here in the global south. Okay, I'm going to just briefly discuss what Code for Africa does and the impact that we work with. So Code for Africa, we're also known as CFA, and we're the continent's largest civic technology organization. Okay, our team is very diverse. We have journalists, data analysts, technologists, and designers distributed all through Africa. Okay, and we do a variety of projects. So for example, we run a whole lot of communities. Many of you may have heard Hacks Hackers. We manage Hacks Hackers in various African countries. We have Wana Data, which is a woman network of data scientists and journalists and analysts, because we're really trying to promote women in tech and journalism in Africa. We also have a very exciting team, our African drone team, and they use drone technology to and use to do satellite imagery and mapping for storytelling. Another exciting program we have is the Census Africa project, and what they do is they build and deploy air quality sensors and collect data through that. We have a fact-checking initiative, which is called Pesachek, that operates in 13 countries. And we also have an iLab team that does forensics analysis, such as social media investigations. And more recently, we started a team called Civic Signal that does media monitoring. So there's a major issue in Kenya and Africa as a whole when it comes to the misuse of government funds. I'm going to take a little deeper look into this to provide more context. So in 2010, the Kenyan government decided to now form a decentralized system, where executive and legislative power is divided amongst the national government and the 47 counties. We are now 10 years plus in, and there's this issue of misuse of funds and little to absolutely no accountability. So a leader will promise, okay, we're going to build 10 schools in five years. At the end of the five years, maybe one school is built or nothing at all. So the Auditor General's report that was published last year, it shows that millions or actually billions of Kenyan shillings is being misused. Now, that money is our money, everyday citizens such as you and I, the tax payers money. The question is, how can we hold them accountable? How can we solve the problem? In this room, everyone is data people. And you will tell me, we will use data. Data is the way forward. But the issue with that is that county data or county budget data is very difficult to access. And if it's available, it's published in a way that an everyday citizen cannot understand it. So what Code for Africa did, we worked on a project called Our County, Our Responsibility. This project aimed at upscaling 56 citizen reporters. So when I say citizen reporters, many of them did not go to journalism school. The everyday citizens who decided, I want to work at a radio station and I want a report, okay? So 56 citizen reporters at 14 radio stations across eight counties in Kenya. So what we had did is we worked with them to now help them report on budget and development issues in their counties. Okay, a part of this project was training. We trained them on fact checking and how to produce fact checks because in Africa right now, misinformation and disinformation is a massive pandemic. We also trained them on budget reporting and contracting for them to understand county budget processes because this is just not straightforward. And of course, my favorite, data journalism. We had to show them how to use data in the stories and the impact that it can have with the audience. And lastly, we trained them on how to now share the newly obtained skills they have with others. We also assisted them with story production. We helped them create radio reports that reported on budget and development indicators on topics in health, agriculture, education and infrastructure. We also created little visualizations that went along with their social media stories on Facebook, Twitter and WhatsApp. Okay, a major part or major component of this project was the development of the PESA Year 2 tool. So PESA Year 2 is a Swahili word. Swahili is Kenya's official language and Swahili is spoken across many different African countries. So when translated into English, PESA Year 2 means our money. PESA Year 2 is a platform that now interactively visualizes development indicators. These visualizations can then further be embedded into stories. Okay, so in Kenya, the county creates what is called the county integrated development plan, which is also known as the CIDP. So what the CIDP does for a five-year period, it now in detail indicates what the county is going to do in terms of development. Sort of like the promises it's going to make for that period. So the data on the CIDP that we have on PESA Year 2 is from 2018 to 2022. And then with this data, the user can now go onto the platform. They can look at the interactive visualizations, which is now easier to read than a PDF. They can see the projects that government has promised them and then take that information, go up to your officials and say, okay, this is the project you promised, what is the status? We've also included data from the 2019 household census and also from the national elections that were held last year. The process, it took about one and a half years to get the project up. First, we had to obtain the CIDP reports. So this was easy to access because we got it online. The problem is it was in PDF reports. So we had to go in and scrape hundreds and hundreds of data tables out of these lengthy reports. Each report probably was two to 300 pages. So one county has one report and we did it for eight counties. And then once we had gotten the data into the spreadsheet, we then needed to brainstorm on how to visualize this data. We spent that time and then looking at the data we had, we had determined the types of charts that we want. So from there, we developed a donor chart, a line chart, a bar chart and a tree map. And these charts were designed specifically for the website. Okay, there after the website interface was designed, the tech team built our website. And then after almost a year of all our hard work was the fun part. All the work of scraping the data and cleaning the data, we got to upload it and create the charts and that was just the best, the highlight of it. Okay, so here is Pesayetu in action. So we work with journalists and they use the visualizations in their stories. For example, this story is on how Homa Bay County is prioritizing healthcare. Okay, and the journalist had taken a visualization on the projects in the healthcare sector and embedded it within their story. So the challenges that we faced with Pesayetu is a similar challenge that we face with many of the projects that we work with. Firstly, data table stuck in lengthy PDF reports and with tools like Tabula, it made it easier. Counties reported data differently, okay? And this was difficult because one county will publish data on the number of children in private schools and another county will publish data on the number of children in high schools. So it makes comparison very difficult and then counties publish data relevant to their development goals making comparison even more difficult. So for example, if a county is very big on tourism, the data or the priorities or the goals they set is for tourism whereas if another county has more agriculture and the income comes from agriculture, the development and goals will be on agriculture. So it's very difficult to now compare projects and budgets amongst these counties. We had issues where the data wasn't adding up. The figures just weren't adding up. So we had to go back, redo the calculations that were given in the table to make sure, okay, we have the accurate data and make sure we find where the error came up from. There was also no metadata attached. So we literally had to go, so for example, if the data had come from like an agriculture department, we had to go to that agriculture department, find their reports and try and find those abbreviations to understand what it meant. Also project related data and the budget data was very robust to visualize because remember you are planning for five years and so the data is broken down by five years with specific projects along the way and it became very difficult. So we had to then brainstorm for hours on end on how to best visualize. Then we had an issue with lengthy data labels. So government departments and project titles, like a project they'll say building of X, Y and Z road or in so and so city, you know, it's very difficult to plot that on a visualization. We then had to call in our in-house copy editing team. So with all these challenges and just with the project on a whole, it wasn't just the data lab that worked on it. We called in our tech team, our designers, our copy editing team. It was an organization-wide project. OK, so if you want to hear more about the work that CFA does, do give us a follow on Twitter. Thanks. That was perfectly on time. That's amazing. Has anyone got any questions for Tricia? You've got one over here. I'll run the microphone. Hi. I wanted to ask about the, did you have to build a scraper for each county? Or did you have all the information of each county centralized somewhere where you could download it? I mean, how manual was the work you had to do? Could you automate most of the project with scraping data from centralized sources? Or did you have to do different scrapers and do a lot of manual work to centralize everything? So each county had its own report, a PDF report. And it was labor intensive, especially how the tables were arranged. Some of the tables will cover like two to three pages. And that now made it quite difficult, and it was quite a manual process, unfortunately, yeah. Any other questions? I have a question. If anyone anywhere else wanted to hold their government to account in this way, what advice would you have for people starting out? I think to start off, it's difficult to do it on your own. Civic Tech, because I'm from Civic Tech, so I'm a big advocate for it. Civic Tech can actually help bridge that gap between government and also civil society, because remember, you also need, so Civic Tech works on building the tool, looking at the data. Civil society will have the call to action, you know, and working together, you can hold government accountable. And also interesting with that is that there's always debate on is government coming to the table, you know? We at Civic Tech, many of us are non-profits. And, you know, we approach government and we're like, we can create a really amazing tool for you. But there's just so much of bureaucracy and red tape around it that it now makes it difficult to work with government. Thank you. Mamal? Thank you for the presentation. You said something really interesting about holding government to account through tracking their promises and having everyday citizens like sort of engage with their governments around, you've promised this, and so what's the progress? I'm just really curious about sort of beyond, like once this is published, is there a way that you're able to track that sort of engagement from everyday citizens or was there some sort of training or effort to show everyday citizens how to best employ this sort of tool? Yeah, so one of our core focuses at Code for Africa is journalism. So we train journalists on using the tools, okay? And like I say to journalists, you know, I'm a data person. I can work with these numbers, but you have a power. You take this data and you make a story about it. This data has meaning, it has a life, it represents someone. By doing so, you give a voice to the people, you inform them, you can hold them accountable. In South Africa as well, where I'm from, you know, journalists are breaking stories when government is still in denial. And then only months later, then would government start to investigate. The same can be said for civil society as well. Yeah, but we do a lot of training on this tool as well. Okay, we've got time for one more question. And if not, oh, Marie. Thank you so much, Tricia, for this excellent presentation. I was wondering about the impact, if you can tell us about how things have changed after journalists using the data, citizens able to read these more friendly data structures. Any good impact stories? So Marissa, this is something I would love to chat with everyone about. Africa is a place, people will use your data, use your tools. But because of limited resources and capacity, they won't come back to you and say, okay, this is what happened or that is what happened, unless we are like monitoring it using Google tools or something like that. And I actually want to have ideas of how can we measure impact. Because journalists, it's so difficult for them to even attend trainings because they're constantly on the field. So to get them to get feedback, to fill in a form, to do anything like that is such a mission. So I would love to actually measure impact in a much better way. And I mean, non-profit, you have to show your funders impact. You get more money here. That's brilliant. And let's remind you there's data tables. You can sign up for the on conference today and tomorrow. If anyone wants to have that discussion or brainstorm together, that's possible. Thank you. Thank you. Thank you for a wonderful talk. Thank you all for your questions.