<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Gemma Ritchie on Medium]]></title>
        <description><![CDATA[Stories by Gemma Ritchie on Medium]]></description>
        <link>https://medium.com/@gejritchie?source=rss-5a9f7c393039------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*seZ8FXPzPqWp5sSfCqj-Ng.jpeg</url>
            <title>Stories by Gemma Ritchie on Medium</title>
            <link>https://medium.com/@gejritchie?source=rss-5a9f7c393039------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 03:58:50 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@gejritchie/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[How a Swahili-language data journalism group created a sustainable business model]]></title>
            <link>https://medium.com/media-hack-magazine/how-a-swahili-language-data-journalism-group-created-a-sustainable-business-model-216f1b2ee3f8?source=rss-5a9f7c393039------2</link>
            <guid isPermaLink="false">https://medium.com/p/216f1b2ee3f8</guid>
            <category><![CDATA[media-hack-collective]]></category>
            <category><![CDATA[sustainable-journalism]]></category>
            <category><![CDATA[nukta-africa]]></category>
            <category><![CDATA[african-journalism]]></category>
            <category><![CDATA[data-journalism]]></category>
            <dc:creator><![CDATA[Gemma Ritchie]]></dc:creator>
            <pubDate>Thu, 16 Feb 2023 21:39:17 GMT</pubDate>
            <atom:updated>2023-02-16T21:39:17.824Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*EsuVmao77g3vpiPb9NaXLw.jpeg" /><figcaption>Left to right: Daniel Mwingira, Maphosa Banduka, Nuzulack Dausen</figcaption></figure><p>When Nukta Africa, the Swahili language data journalism news site based in Tanzania, launched in 2018 the founders quickly realised that to be sustainable, news alone wasn’t going to cut it. Just running, as they put it, “strange stuff about data” didn’t generate enough revenue.</p><p>Instead of walking away, founders and journalists <a href="https://tz.linkedin.com/in/nuzulack-dausen-76966980">Nuzulack Dausen</a>, <a href="https://tz.linkedin.com/in/maphosa-banduka-343440185">Maphosa Banduka</a> and <a href="https://tz.linkedin.com/in/daniel-mwingira-772a7422">Daniel Mwingira</a> went hybrid. They began offering their data <a href="https://nuktaafrica.co.tz/">services</a> to other media organisations, companies, NGOs, and people who wanted to learn more about data.</p><p>Now, in addition to the data <a href="https://nukta.co.tz/">news site</a>, they have created a <a href="https://www.youtube.com/@NuktaTV/about">TV show,</a> a <a href="https://www.facebook.com/NuktaFakti">Facebook site</a> dedicated to fact-checking, a <a href="https://podcasts.apple.com/za/podcast/nukta-the-podcast/id1573241582">podcast</a>, and <a href="https://jikopoint.co.tz/">Jiko Point</a>, a multimedia platform that focuses on clean energy. And they’ve gone from a team of three to 13.</p><p>Media Hack spoke with the Dausen via email. This is an edited version of our conversation.</p><p><strong>What went into your decision about producing data journalism in Swahili?</strong></p><p>When we were starting our company, data journalism was at the nascent stage in Tanzania. And it still is. Even the news outlets that were publishing data-driven news were mainly English newspapers or websites. Most Swahili online news outlets were publishing breaking news, gossip, or entertainment. Since launching, we have influenced many mainstream and digital-native news outlets to publish data-driven content in Swahili.</p><p><strong>How has your work been received?</strong></p><p>While we are still growing our audience, the feedback we have received is encouraging. For example, in 2021 we wanted to close a data product that presents wholesale crop prices three times a week. The analytics behind the product were not impressive. Within a week of stopping it, we received many calls from readers from different parts of the country asking why we were not publishing the crop data. They told us the data was useful to them. We reinstated the product and it is still there today.</p><p><strong>What is Nukta’s business model?</strong></p><p>We use a hybrid business model as part of our strategy to diversify our revenue sources. Our revenue is generated from advertisements; sponsored content; affiliate marketing; grants; training and consultancy services; and content productions services where clients pay us to create content to be published on their platforms, including data visualisations.</p><p>We started experimenting with a membership model for our special reports for both Nukta Habari and <a href="https://jikopoint.co.tz/">Jiko Point</a> in January 2023, and we expect by December will have some insights on how it has been received.</p><p><strong>What are the barriers to data literacy you have found during your data journalism training?</strong></p><p>The majority of trainees do not have basic digital literacy skills that would help them quickly adopt data journalism. This means most of the time in data journalism courses we are teaching digital tools instead of actual data-related stuff.</p><p>Also, there is a belief that if someone failed mathematics in secondary school, they won’t make it when he or she has a job. This belief has made the majority of our trainees believe data journalism is very difficult, while it is not.</p><p>In some localities, it is difficult to access data, and top managers in media outlets don’t understand all the benefits of data journalism. Some have either rejected data journalism programs or do not support journalists taking more time to do just one data-driven story.</p><p>To solve this, we need to customise data journalism training based on the level of understanding by putting the courses in ranks such as basics, intermediate, and advanced.</p><p>Also, the training should consider resources available for respective trainees…we need to invest more in basic digital skills such as the use of Microsoft Excel, social media, and cloud-based services like Google Drive.</p><p><strong>If you could change one thing about data journalism, what would it be?</strong></p><p>We have witnessed many journalists attending several data journalism training courses which include data analysis and data visualisation. But they are unable to write or produce quality data-driven stories. The main issue is storytelling. To solve this problem, we are advocating dedicating more hours to how to write data stories or scripts.</p><p>Also, we would like to change the perception among people who believe data journalism is only for digital news outlets and newspapers. We started working on this by ensuring we find innovative ways of telling data-driven stories on radio, like the simplification of percentages in narrations and adding more human voices.</p><p>Also, we need to help journalists and editors in the contextualisation of stories. Most of them just drop the findings in stories without giving the context which will answer the so-what factor. This downplays the impact of the stories.</p><p>Our stories and infographics or visuals should strive to contextualise the issue in question. If the food prices are rising in Tanzania, why should we care? Who is affected the most? What can be done to solve that? A series of visuals can bring a story context.</p><p><strong>What is the difference between collecting and working with data in the Global North versus the Global South?</strong></p><p>The main difference is how we value data in our daily lives and business. I have been closely following the open data movement and data journalism progress across the world, and most literature on the collecting and use of data are from Europe and America. At least we are seeing some work from Asia but very rarely from Africa.</p><p>Governments and other institutions in the Global North are at least providing data and the users have a relatively high data literacy rate.</p><p>In the Global South, while we have improved in data collection and use within the last decade, we still need to do more. There are times you find government websites have old data, probably three to five years old. Some countries do not have data protection and access to information laws which are key in data collection and use.</p><p><strong>Do you have any tips for data journalists working on the continent, or those interested in pursuing data journalism?</strong></p><ol><li><strong>Never stop learning</strong> new things in data journalism. Try new tools, ways of storytelling, and data analysis and visualisation. Things change constantly.</li><li><strong>Generate and understand your sources of data.</strong> A good data journalist knows many data sources and ways of getting them. For websites or data portals, turn on notifications so you don’t miss anything.</li><li><strong>Understand the nature of the data you are working with.</strong> Read the metadata;<strong> </strong>it gives you details on who published the data, how it was collected, what measurements are used, and how frequently it is published. For example, some data refers to households as units of analysis but someone who fails to understand his or her dataset well will be likely to say people instead of household. This is wrong. Take time to check everything.</li><li><strong>Use simple language and simplify the percentages.</strong> We are doing data journalism not to brag to our audience that we know how to work with data — we do this to inform them and help them make informed decisions.</li><li><strong>Practice. Practice. Practice.</strong> Immediately after attending a training session, practice what you have learned. We have trained thousands of journalists and editors, but very few continue to do data journalism well. No miracle here. You will improve by learning from past works and getting feedback about them.</li></ol><p><strong><em>Media Hack Collective offers </em></strong><a href="https://mediahack.co.za/#/training"><strong><em>data journalism training</em></strong></a><strong><em> across the African continent and publishes data journalism on </em></strong><a href="https://theoutlier.co.za/"><strong><em>The Outlier</em></strong></a><strong><em>, including a growing </em></strong><a href="https://charts.theoutlier.co.za/"><strong><em>collection of charts</em></strong></a><strong><em> and </em></strong><a href="https://newsletters.mediahack.co.za/"><strong><em>data-backed newsletters</em></strong></a><strong><em>.</em></strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=216f1b2ee3f8" width="1" height="1" alt=""><hr><p><a href="https://medium.com/media-hack-magazine/how-a-swahili-language-data-journalism-group-created-a-sustainable-business-model-216f1b2ee3f8">How a Swahili-language data journalism group created a sustainable business model</a> was originally published in <a href="https://medium.com/media-hack-magazine">Media Hack</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to use InVID to pull together tweets using a hashtag and a specific time period]]></title>
            <link>https://medium.com/media-hack-magazine/how-to-use-invid-to-pull-together-tweets-using-a-hashtag-and-a-specific-time-period-b65883b0c65a?source=rss-5a9f7c393039------2</link>
            <guid isPermaLink="false">https://medium.com/p/b65883b0c65a</guid>
            <category><![CDATA[media-hack-collective]]></category>
            <category><![CDATA[how-to]]></category>
            <category><![CDATA[tools-for-data]]></category>
            <category><![CDATA[data-journalism]]></category>
            <category><![CDATA[african-journalism]]></category>
            <dc:creator><![CDATA[Gemma Ritchie]]></dc:creator>
            <pubDate>Thu, 17 Nov 2022 08:44:24 GMT</pubDate>
            <atom:updated>2022-11-17T08:44:24.536Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/820/1*GZ-WRsAY3KK-AJM_pMX2JA.png" /></figure><p>The InVID project — funded by the <a href="https://www.invid-project.eu/description/">European Union’s Horizon 2020 research and innovation programme</a> — is a powerful tool to help journalists, researchers, and fact checkers pull tweets from a specific hashtag during a set period of time, which is great for figuring out a timeline of events after they happened.</p><p>The tool can also help with analysing the doctoring of video and photographs but we’re just working with the tweet abilities here.</p><p>Please note that to use InVID’s Twitter social network analysis (beta), you need to be authorised to use the tool and acknowledged as a working journalist, researcher, or fact checker.</p><p>We’ll focus on the climate change conference (COP27) in Egypt from 6 November to 18 November to give you an example of how to use it.</p><h3>Setting it up</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/314/1*ub-saotnvjd0VWxR8cHMmg.png" /></figure><p>Start by installing the <a href="https://www.invid-project.eu/tools-and-services/invid-verification-plugin/">InVID extension</a> on your Chrome Browser. To launch the extension, go to the puzzle icon on the top right-hand corner of your browser screen.</p><p>Once you have launched it, log in for authorisation with your email. You will be sent an alpha-numeric code within a few minutes which you will use to log in.</p><p>Now, let’s start digging</p><p>You will find the Twitter social network analysis under the “Data Analysis” tab in InVID.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/550/1*E8BZq-0PaLn0u_I6e2jPSQ.png" /></figure><p>Once you have selected Twitter SNA, you will be taken to a page where you can fill in your search parameters.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/550/1*3_t-oiPKYfMjJkICnm7Qdg.png" /></figure><p>After you have placed your parameters, InVID will pull the tweets for that hashtag and time period. You can also add additional parameters such as selecting a language, including pictures or videos, and/or excluding some words.</p><p>InVID will pull the tweets for you using your hashtag and time parameters in a couple of minutes.</p><p>Once it has pulled the data, scroll down to the “8. Heatmap of tweets distribution”. We will use this graphic to show you how to pull the tweets.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/550/1*q03oFakmSAt0BDrr4QjVzQ.png" /></figure><p>Now we are going to extract our tweets.</p><p>To do this, run your cursor on the blocks in the heatmap to find out if there were tweets using your criteria at that time. Each block in the heatmap represents an hour.</p><p>Once you have clicked on the block, the tweets associated with it will appear below the graph. And in the right-hand corner, there will be an option to download the tweets.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/550/1*srvIZgTPSvNEC2-aSqMsmA.png" /></figure><p>The tweets will download in a .CSV (comma-separated values) file which can be opened in Google Sheets, which is a free tool.</p><p>Then you will have the following columns:</p><ul><li>twittersna_result_date (the date and the time of the tweet);</li><li>twittersna_result_username (the username);</li><li>twittersna_result_tweet (the text in the tweet);</li><li>twittersna_result_like_nb (the number of likes the tweet received);</li><li>twittersna_result_retweet_nb (the number of retweets);</li><li>elastic_url (the link to the tweet).</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/550/1*MtmcFQjpr4u6sYyk0H_tXA.png" /></figure><p>Now, all you need to do is repeat the steps for each block on the heatmap, and you will have your data.</p><p>If you want to join the datasets together, you can do this in Google Sheets, but the bigger a data set gets, it might be easier to join it using programmes like <a href="https://posit.co/download/rstudio-desktop/">R Studio.</a></p><p>Check out our article on <a href="https://inside.mediahack.co.za/five-reasons-google-sheets-is-better-than-microsoft-excel-d36f43f21dea"><strong>Five reasons Google Sheets is better than Microsoft Excel</strong></a><strong>. </strong>For quick videos on what you can do with Google Sheets, check out our <a href="https://www.tiktok.com/@mediahackza">TikTok channel</a>.</p><p><a href="https://mediahack.co.za/"><strong><em>Media Hack Collective’s</em></strong></a><strong><em> </em></strong><a href="https://mediahack.co.za/"><strong><em>The Outlier</em></strong></a><strong><em> publishes a data journalism newsletter every two weeks. Read the latest issue and subscribe </em></strong><a href="https://newsletter.theoutlier.co.za/"><strong><em>here</em></strong></a><strong><em>.</em></strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b65883b0c65a" width="1" height="1" alt=""><hr><p><a href="https://medium.com/media-hack-magazine/how-to-use-invid-to-pull-together-tweets-using-a-hashtag-and-a-specific-time-period-b65883b0c65a">How to use InVID to pull together tweets using a hashtag and a specific time period</a> was originally published in <a href="https://medium.com/media-hack-magazine">Media Hack</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Talking data journalism with Goodman Lepota]]></title>
            <link>https://medium.com/media-hack-magazine/talking-data-journalism-with-goodman-lepota-b311b2435975?source=rss-5a9f7c393039------2</link>
            <guid isPermaLink="false">https://medium.com/p/b311b2435975</guid>
            <category><![CDATA[journalism]]></category>
            <category><![CDATA[africa]]></category>
            <category><![CDATA[media-hack-collective]]></category>
            <category><![CDATA[south-africa]]></category>
            <category><![CDATA[data-journalism]]></category>
            <dc:creator><![CDATA[Gemma Ritchie]]></dc:creator>
            <pubDate>Thu, 13 Oct 2022 19:44:07 GMT</pubDate>
            <atom:updated>2022-10-13T19:51:10.271Z</atom:updated>
            <content:encoded><![CDATA[<p>Goodman Lepota has loved telling stories with data since he was a teenager. When he was 16, he started a newspaper at his high school in Tembisa, just north of Johannesburg, South Africa.</p><p>Now a <a href="https://www.linkedin.com/in/goodmanlepota/?originalSubdomain=ie">Google Account Manager</a> based in Dublin, Ireland, he launched <a href="https://www.instagram.com/preamble.africa/">Preamble</a> in <a href="https://www.preamble.africa/about">May 2022</a>. His data visualisations about Africa are published on <a href="https://www.instagram.com/preambleafrica/">Instagram </a>and <a href="https://www.preamble.africa/subscribe/">Substack</a>. Here is an edited version of our conversation.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*EUqUkB-OyCk0Lw40tQR6wQ.png" /><figcaption>Goodman Lepota, the founder of Preamble</figcaption></figure><p><strong>When did you become interested in data?</strong></p><p>When I was younger, my mom sold second-hand clothing to families in our community. When she restocked her inventory, she would buy me used <em>National Geographic </em>and<em> TIME</em> magazines, and encyclopedias from antique shops. My fascination with data and visualisations began there.</p><p><strong>When did you first start using data in journalism?</strong></p><p>In eighth grade, I started a school newspaper with the goal of telling local stories using data. When I earned a scholarship from Jiyana Secondary School to the boarding school African Leadership Academy, I founded a media company for students. Then everything changed when I moved from South Africa to Marist College in New York, US. I joined the university newspaper and became the publisher and creative director.</p><p>Preamble embodies a passion I always had that I hadn’t explored fully. With Preamble, I got to embrace telling stories with data.</p><p><strong>Why did you decide to start Preamble?</strong></p><p>I read a lot of international news and I always have ideas about things I am curious about. There is something I call the “Africa Data Gap.” I read a story from the <em>Financial Times </em>or<em> Bloomberg </em>and find the analysis covers the Americas, Europe, the Middle East, the UK and Asia Pacific, while data on Africa is sometimes unavailable. I think to myself. “Wow, what would a dataset on Africa look like?”</p><p><a href="https://www.preamble.africa/archive">Preamble newsletter archives</a></p><p><strong>When do you work on Preamble?</strong></p><p>I do most of the data analysis and visualisation after work or on weekends. My friends joke about it now when I have to go home and ask, “Oh are you going home to make graphs?”</p><p><strong>What is important for readers to know about data journalism?</strong></p><p>Data journalism influences decisions. That is why Preamble is “For decision makers making an impact in Africa.” In every decision that is made, data is used. While living in the US I discovered that some of the decisions that governments made were influenced by data journalism; if the <em>New York Times</em> broke a data story, you could trace from when that story broke to discussions about legislation or policy in the Senate.</p><p><strong>If you could change one thing about data journalism, what would it be?</strong></p><p>International news organizations need to include Africa in their analysis. In Africa, I would say newsrooms need to focus on data journalism.</p><p>Overall, it is about simplification. It is something I think a lot about. Most of the data I see is presented in a very interesting and compelling way. But when I look at it, I think one would need to how to read the graphs to understand it. I would argue that some graphs tend to be created by reporters in news organisations speaking to other journalists. With Preamble, I think about simplification a lot. I think: How do I make this graph accessible to a person who isn’t into data viz or a data journalist?</p><p><strong>What is your favourite data tool?</strong></p><p>I use Excel the most, and Photoshop and Illustrator to visualise data. Most of my visualisations are designed from scratch.</p><p><strong>What do you think are the major differences in telling data stories in the West when compared to the continent or the broader Global South?</strong></p><p>I am obsessed with things that are designed well. On the continent, newsroom data presentation could be improved. For instance, I was in Nairobi, Kenya, where I picked up a business newspaper because I wanted to see how it visualised data. Most of it was cartoons. I thought, “Hmm, I would have actually presented this in a more flat way.” I would have removed the cartoons and used icons to make the data look more sophisticated and clean.</p><p>If I had to change anything, and were to consult newsrooms and governments, I would present the data in the most beautiful way.</p><p><strong>Do you have any tips for people wanting to start analysing data?</strong></p><p>If there is anything I could learn but don’t have time for it, it is D3. Data visualisation would change so much on the continent if we used D3.</p><p><strong>What sorts of data projects would you like to see more of on the continent?</strong></p><p>There are so many data projects I want to see. If there is anything that wakes me up at night and I am like “wow” — it is the <a href="https://www.nytimes.com/interactive/2018/03/19/upshot/race-class-white-and-black-men.html"><em>New York Times</em>’ article on neighbourhoods</a>. It is a story connected to me. I grew up in the township, which is a low-income area. I have always been very curious: How is it that I ended up where I ended up? Could a data story explain that? There is a part of me that wants to do a research project about the kids that grew up in that area.</p><p>The data part of me is thinking, we lived one street from a library. In our high school, there was a library and outside of school there was another library, which means there were three libraries in our proximity. Then there were five schools near my house. Also, there were other people I know from the neighborhood who started really successful businesses.</p><p>I wonder if there is something related to data that can explain how we came from this rough part of the township to do what we did. Is it maybe the proximity of the library? How we were raised? Or the time in which we were raised? But I think that story is missing in South Africa. If I were in the South African government, I would do a trace study on neighbourhoods and how people are growing up, then do the impact story in the later years and reveal whether the policies have had an impact.</p><p><strong>Do you call yourself a data journalist?</strong></p><p>I would say no, because journalism requires a human impact or human story. I think data journalists present that. I would say I am really passionate about telling stories in a compelling way. But if you look at my newsletters, they miss the human impact story. So, I don’t think I qualify. But there is something for me to do on the continent, like collaborate with journalists. I want to bridge the data gap on the continent. I want to read something online that actually includes Africa. But I don’t think the world is there yet. We cannot talk about prosperity and the future without closing the data gap.</p><p>Media Hack Collective publishes data journalism on <a href="https://theoutlier.co.za/"><em>The Outlier</em></a><em> </em>— including a growing <a href="https://charts.theoutlier.co.za/">collection of charts</a> — offers <a href="https://mediahack.co.za/#/training">data journalism training</a> across the African continent, and publishes a <a href="https://newsletters.mediahack.co.za/">number of data-backed newsletters</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b311b2435975" width="1" height="1" alt=""><hr><p><a href="https://medium.com/media-hack-magazine/talking-data-journalism-with-goodman-lepota-b311b2435975">Talking data journalism with Goodman Lepota</a> was originally published in <a href="https://medium.com/media-hack-magazine">Media Hack</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Finding, tracking and organising datasets isn’t as difficult as you might think. Here’s our guide.]]></title>
            <link>https://medium.com/media-hack-magazine/finding-tracking-and-organising-datasets-isnt-as-difficult-as-you-might-think-here-s-our-guide-46bf4cf26737?source=rss-5a9f7c393039------2</link>
            <guid isPermaLink="false">https://medium.com/p/46bf4cf26737</guid>
            <category><![CDATA[media-hack-collective]]></category>
            <category><![CDATA[data-journalism]]></category>
            <category><![CDATA[medium]]></category>
            <category><![CDATA[how-to]]></category>
            <category><![CDATA[journalism]]></category>
            <dc:creator><![CDATA[Gemma Ritchie]]></dc:creator>
            <pubDate>Thu, 22 Sep 2022 20:38:01 GMT</pubDate>
            <atom:updated>2022-09-23T05:27:10.852Z</atom:updated>
            <content:encoded><![CDATA[<h3>Finding, tracking and organising datasets isn’t as difficult as you might think. Here’s our simple guide.</h3><p>Data isn’t always readily available in a clean dataset waiting to be analysed and visualised. But that’s not always the case.</p><p>Sometimes there is no data. Or there is information available but not in a spreadsheet, such as academic papers, government white papers or statements in PDF. So, here are some steps to build a dataset so you don’t (hopefully) end up pulling your hair out.</p><p>You will need:</p><ul><li>a hypothesis you want to test;</li><li>a spreadsheet (Google Sheets or Excel);</li><li>a variable (or a set of variables) you want to track;</li><li>a means to keep your dataset up-to-date (news alerts, newsletters, social media, press releases).</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4IFvCqcJuJpXe00nzFcM5g.png" /></figure><h4><strong>A hypothesis</strong></h4><p>I wanted to find out what was happening in by-elections in South Africa following the local government elections on 1 November 2021, and see whether by-elections were the result of ward councillors dying or resigning from their post. I also wanted to see if one of South Africa’s nine provinces had more by-elections than others.</p><p>The Independent Electoral Commission (IEC) supplied data on when the by-election happened, who the previous incumbent was and who had replaced them in each ward.</p><p>But the dataset supplied by the IEC didn’t explain why the by-elections took place. However, a few weeks before the by-elections, the IEC would share the reasons for them— without naming the ward councillor — in their press releases. These reasons included death, resignations, political parties terminating memberships and court orders — as in when a court set aside the election results.</p><p>I knew what information I wanted to track and where to find it. Now, I needed to organise the information into a dataset to track it over time.</p><h4>A spreadsheet</h4><p>There are several ways to keep track of information. My favourite is <a href="https://inside.mediahack.co.za/five-reasons-google-sheets-is-better-than-microsoft-excel-d36f43f21dea">Google Sheets</a>. It is free and great for collaboration. You can use these same ideas with Microsoft Excel or if you are old-school pen-and-paper.</p><p>In a clean spreadsheet, I created several columns to answer the questions from my hypothesis using the journalist crib list: when something happened, where it happened, what happened and why it happened and who was involved.</p><h4>A set of variables</h4><p>My first column was my date column: “when”. In this column, I noted the date of the by-election, which is important for fact-checking and keeping your data organised.</p><p>Then I had my three “where” columns, which answered where the by-election took place, including the province, the municipality and its unique alpha-numeric code. The alpha-numeric code, though not necessary for general visualisations, like bar charts, is useful if you <a href="https://inside.mediahack.co.za/how-to-make-custom-maps-even-if-you-dont-know-how-to-code-af3274bfe0a4">need to make maps</a> and merge your dataset with GeoJson files.</p><p>The what column came next. In my first “what” column, I had a column called “brief description” where I described the events that happened — as whether the by-election was won or lost by the incumbent party and the reason for the by-election.</p><p>Then came the “who” columns. Because I wanted to track the incoming parties of the ward councillors, I had the columns “new appointee”, “new party”, “previous party”, and “previous appointee”. I kept track of the ward councillor appointees as the data was available because it made it easier to monitor if a ward councillor changed parties.</p><p>And then, to summarise the data I was collecting on whether the party stayed in power or not, I had a column called “party retained”, which I filled in with either yes or no.</p><p>But to write a story you have to contextualise an issue. People always want to know the reason — the why — something happened. In this case, why did by-elections happen less than a year following the 2021 local government elections?</p><p>To answer this, I added a column called “reason for by-election”. It was one or two-word summaries, such as death, resignation, court order or party removal. In the case of death, since not all the ward councillors died from natural causes, as I found while researching my <a href="https://theoutlier.co.za/elections/83004/16-out-of-south-africas-28-by-elections-held-this-year-following-ward-councillors-deaths">story</a>, I added the column “reasons for death”, which included the variables: murder (shot at leaving the office or home, or while driving), illness (either sudden or a long battle), car crash and unknown (where there is no information in news articles, social media posts or from statements by the party on the nature of the councillor’s death).</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6YOuKUV347rWrf8myHssOw.png" /></figure><p>I also had several columns called “source” where I placed links to the information I found. These columns are crucial resources if you need to double-check the data you put into your dataset.</p><p>One of the things I learned from making this dataset is that you need to keep your variable names the same throughout the process. Otherwise, it will make it difficult to analyse your data. I made a tab called “about this data”, which helped me remember what details went into my columns and how I phrased variables in my dataset. This will also make it easier for others to use your dataset.</p><p>If you want to see a part of the dataset I have been working on, see the <a href="https://theoutlier.co.za/news/82348/sa-municipality-tracker">Municipality Tracker</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*oMprHQDDXNAstOolgF2oRA.png" /><figcaption>An example of what the spreadsheet looks like</figcaption></figure><h4>A means to keep up-to-date</h4><p>You now have your dataset set up. How do you keep it up-to-date? One of the easiest ways to keep track of information is to set up a <a href="https://www.google.com/alerts">Google Alert</a>. It will send you web pages, news articles and video links when your keywords are identified. You can have it send you an update once a day, a week or each time it sees an update. Preferably, go for the once-a-day or once-a-week option; otherwise, you might overwhelm your inbox.</p><p>The second option is to monitor certain accounts using Twitter lists which you can check throughout the day using Twitter or <a href="https://tweetdeck.twitter.com/">Tweetdeck</a>. If you have some coding skills, you can put together a <a href="https://developer.twitter.com/en/docs/tutorials">Twitter scraper</a>, or if you have a budget, you can pay for a Twitter scraper application like <a href="https://workspace.google.com/marketplace/app/tweet_archiver/976886281542">Tweet Archiver</a>, a Google Sheets Extension which is easy to use and costs $4.95 a month. But for the most part, the Google Alerts are effective.</p><p>Setting up the structure for your dataset shouldn’t take long, and it is fluid. Over time, you may add more columns. Just keep your information organised so you can use it quickly and painlessly!</p><p>So why did South Africa have 28 by-elections by July? Most of the by-elections were because a ward councillor had died. You can read my full story here: <a href="https://theoutlier.co.za/elections/83004/16-out-of-south-africas-28-by-elections-held-this-year-following-ward-councillors-deaths">Two-thirds of by-elections since November 2021 were because a councillor died</a>.</p><p><em>Note, after publishing, this piece was lightly edited for clarity.</em></p><p><a href="https://mediahack.co.za/"><strong><em>Media Hack Collective</em></strong></a><strong><em>’s</em> </strong><a href="https://theoutlier.co.za/"><strong>The Outlier</strong></a><strong> <em>publishes a data journalism newsletter every two weeks. Read the latest issue and subscribe </em></strong><a href="https://newsletter.theoutlier.co.za/"><strong><em>here</em></strong></a><strong><em>.</em></strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=46bf4cf26737" width="1" height="1" alt=""><hr><p><a href="https://medium.com/media-hack-magazine/finding-tracking-and-organising-datasets-isnt-as-difficult-as-you-might-think-here-s-our-guide-46bf4cf26737">Finding, tracking and organising datasets isn’t as difficult as you might think. Here’s our guide.</a> was originally published in <a href="https://medium.com/media-hack-magazine">Media Hack</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to make custom maps — even if you don’t know how to code]]></title>
            <link>https://medium.com/media-hack-magazine/how-to-make-custom-maps-even-if-you-dont-know-how-to-code-af3274bfe0a4?source=rss-5a9f7c393039------2</link>
            <guid isPermaLink="false">https://medium.com/p/af3274bfe0a4</guid>
            <category><![CDATA[flourish]]></category>
            <category><![CDATA[how-to]]></category>
            <category><![CDATA[media-hack-collective]]></category>
            <category><![CDATA[data-journalism]]></category>
            <category><![CDATA[maps]]></category>
            <dc:creator><![CDATA[Gemma Ritchie]]></dc:creator>
            <pubDate>Thu, 07 Jul 2022 18:14:34 GMT</pubDate>
            <atom:updated>2022-07-07T18:54:37.857Z</atom:updated>
            <content:encoded><![CDATA[<h3>How to make custom maps — even if you don’t know how to code</h3><p>When programmes like <a href="https://flourish.studio/">Flourish</a> or <a href="https://www.datawrapper.de/">Datawrapper</a> don’t have the map format you need — and you don’t know how to code — you can still make a custom map. And it’s actually pretty easy.</p><p>All you need are:</p><ul><li>A file with your <a href="https://data.humdata.org/dataset">geodata</a></li><li><a href="https://mapshaper.org/">Mapshaper.org</a></li><li><a href="https://flourish.studio/">Flourish</a>, if you need to merge data onto your map</li><li>Or, if you just need a map image without merging data onto it, you can use Figma, Adobe Illustrator, Adobe XD or Photoshop to add colour</li></ul><p>First, work out what kind of map you want to make.</p><p>In my case, I needed a map of South Africa showing the local municipalities so I could mark which municipalities were run by a particular political party.</p><h3>Getting the data</h3><p>I used the UN’s OCHA <a href="http://data.humdata.org/">data.Humdata.org</a> and found the geodata set “<a href="https://data.humdata.org/dataset/cod-ab-zaf">Subnational Administrative Boundaries</a>”. According to its description, this geodata had “South Africa administrative levels 0 (country), 1 (province), 2 (district), and 3 (local municipality) boundaries.” Under data and resources, I downloaded the zipped file “zaf_adm_sadb_ocha_20201109_SHP.zip”<strong>.</strong></p><p>Here are a few things to think about when picking your geodata:</p><ul><li>Not all map data is equal, so be sure to check when your map was last updated — especially in the run-up to elections.</li><li>Maps require several elements, in the form of files. So don’t be afraid when you have to import four or five files for the next step.</li><li>If you are unsure which zip file is correct, use the following steps to visualise the map quickly in mapshaper.org to make sure you have the right data.</li></ul><h3>Making the map</h3><p>In our zip file, there are data for the country, provinces, district and local boundaries and a series of files that contains all of these boundaries.</p><p>So, let’s look for the file that has the information we need. In our case, the number 3 corresponds to data for local municipality boundaries so we chose the file with “adm3” in the filename. Once you identify the files you need— these will end with .shp, .shx, .sbx, .sbn, .prj, .dbf, .cpg — drag them into <a href="http://mapshaper.org/">mapshaper.org</a> and choose the option to “import all”.</p><p>You’ll know quickly if you are importing the wrong file or are missing one or two. You will either have no map in <a href="http://mapshaper.org/">mapshaper.org</a> or, for example, if you just import the shape file you will have the map&#39;s geometry, but you won’t have any of the underlying data to help you merge the map with a dataset.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hjgRnEjIfw991jEt-9GsBw.png" /></figure><p>With the correct files imported, you’ll have your map. But it might be pretty big because of the level of detail, so you can simplify it in <a href="http://mapshaper.org/">mapshaper.org</a> to as small as 10% or less by using the simplify tool.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uUPR3QZnPRXyod6wtOPLdA.png" /></figure><p>This is my experience working with the South African map. In other maps, the amount you can simplify — before it becomes unrecognisable — may vary quite a bit. If you simplify the map to zero, for example, you will have an abstract simplification of your map.</p><p>Now you can download your map as a .svg and begin working with it in design programmes like Figma or Adobe Illustrator.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*N0jG0XSyxkoALW-aXPY1Xw.png" /></figure><p>But if you want to merge a lot of data onto the map, download it as a geo.JSON file and head over to Flourish.</p><h3>Using Flourish</h3><p>In Flourish, look for the “Blank Map” option. Once you have selected it, drag your downloaded .geoJson file onto Flourish’s “Data” section.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*WxiZJZhmQL3tvIyG5J8PFw.png" /></figure><p>You’ll find two tabs: Regions and Points.</p><p>If you are making a choropleth map (where regions/districts/municipalities are in certain colours to visualise an idea), you only need to worry about the Regions tab when you upload your map. But if you want bubbles (or other shapes) to visualise, say, population density or the number of children in schools, then you will need to use the Points tab.</p><p>We’ll make a <strong>choropleth</strong> map first.</p><p>Drag your data in .csv or .xls into Flourish and select the option to “Merge”.</p><p>Flourish is pretty amazing and will automatically select the columns most likely to match. In this case, when my team and I made the dataset with the governing party in each municipality (over 200 rows of data) we not only had the name of the municipality but the municipal code, which is a combination of letters and numbers. (Flourish also gives you the option to overwrite its suggestions and instead merge the data on two columns that you chose from each dataset.)</p><p>In this same “Merge” dialogue box, you will also get several other options to select from. These options ask whether you want to keep the unmatched data from the current dataset and merge it without a match from your incoming dataset. I like to keep all the data as that will tell me pretty quickly if I made a mistake like merging the dataset with the wrong columns or if there is missing data or misspelt variables.</p><p>In our example, Flourish used the municipal code from the map data and from our own data to join the two datasets.</p><p>So now you have your map and your merged data. But in the preview, your map may look pretty small. This is because Flourish’s automatic setting is to frame the map in a world view. If you change the map to “Auto”, Flourish will reframe your map to fit the border around it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uJRqL9iPkr111jQuAJHH0Q.png" /></figure><p>Now you can adjust the colours of your regions and add a heading, sub-heading and footer in the “Preview” part of Flourish.</p><p>We even added a drop shadow, which can be found as an option in the “Regions Layer”.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*FlyXIagnSVxG9A_lP3pAHQ.png" /></figure><p>If you want to make a map with <strong>bubbles</strong>, you will need to use the Points tab instead.</p><p>When we used the Regions tab, we needed columns with identical pointers that could be joined along the column named “municipal code”. This time you need to make sure you have latitude and longitude coordinates in separate columns.</p><p>Like we did in the previous example, you will download your .csv but, this time, overwrite your data in the Points tab.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JCRah2SMvJUlVwMAdIgQ1g.png" /></figure><p>When it comes to editing your bubbles map, you need to look at your “Points Layer” in the “Preview” section of Flourish.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*KgkPPplOC5uDYp7bVzaK0Q.png" /></figure><p>In my case, I didn’t need to have different size variables, I just needed different colours but the same sized dot, so I only edited the “Maximum Radius” section of the “Points Layer”.</p><p>When you are finished making a map in Flourish, use the “Export &amp; Publish” button in the top right-hand corner to either download your map as a .svg, .jpg or .png or to make it into a shareable link.</p><p>You can see how we used our own maps in Flourish in “<a href="https://theoutlier.co.za/elections/82042/whos-ruling-who-an-interactive-map-of-the-parties-running-south-africas-municipalities">Who’s ruling who?</a>”, “<a href="https://theoutlier.co.za/news/82813/how-many-bodyguards-does-a-mayor-need">How many bodyguards does a mayor need?</a>” and <a href="https://charts.theoutlier.co.za/chart-view/82882">247 post offices closing in four years</a>.</p><p>If you enjoyed this article, you can find other tips on how to visualise data <a href="https://inside.mediahack.co.za/tagged/tools">here</a>.</p><p><a href="https://mediahack.co.za/"><strong><em>Media Hack Collective</em></strong></a><strong><em>’s</em> </strong><a href="https://theoutlier.co.za/"><strong>The Outlier</strong></a><strong> <em>publishes a data journalism newsletter every two weeks. Read the latest issue and subscribe </em></strong><a href="https://newsletter.theoutlier.co.za/"><strong><em>here</em></strong></a><strong><em>.</em></strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=af3274bfe0a4" width="1" height="1" alt=""><hr><p><a href="https://medium.com/media-hack-magazine/how-to-make-custom-maps-even-if-you-dont-know-how-to-code-af3274bfe0a4">How to make custom maps — even if you don’t know how to code</a> was originally published in <a href="https://medium.com/media-hack-magazine">Media Hack</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Our data on vaccine deliveries to Africa was too much for our map to handle.]]></title>
            <link>https://medium.com/media-hack-magazine/our-data-on-vaccine-deliveries-to-africa-was-too-much-for-our-map-to-handle-cd82be6a97ba?source=rss-5a9f7c393039------2</link>
            <guid isPermaLink="false">https://medium.com/p/cd82be6a97ba</guid>
            <category><![CDATA[mapping]]></category>
            <category><![CDATA[covid19]]></category>
            <category><![CDATA[covax]]></category>
            <category><![CDATA[how-to]]></category>
            <category><![CDATA[africa]]></category>
            <dc:creator><![CDATA[Gemma Ritchie]]></dc:creator>
            <pubDate>Thu, 21 Apr 2022 21:09:50 GMT</pubDate>
            <atom:updated>2022-04-21T21:09:50.599Z</atom:updated>
            <content:encoded><![CDATA[<h3>Our data on vaccine deliveries to Africa was too much for our map to handle. So we changed our strategy</h3><p>The first vaccines for Covid-19, delivered by the vaccine alliance Covax, arrived on the African continent at the end of February 2021, and then in dribs and drabs over the next year.</p><p>Europe, North America, and parts of Asia and the Middle East already started vaccinating their populations by then. As soon as the vaccines finally arrived on the continent, we started gathering the data.</p><p>Our early visualisations began with a map of Africa from <a href="https://www.datawrapper.de/">Datawrapper</a> and a dataset with only eight variables, representing the few countries that had received vaccines. We used <a href="https://academy.datawrapper.de/article/114-how-to-create-a-symbol-map-in-datawrapper">Datawrapper’s symbol map</a>, which allowed the indicator circle to increase in size with the number of vaccines delivered by Covax.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1jgp9wXSYjCxHBZt4D9yjA.png" /></figure><p>We used this method for the majority of 2021. But, eventually, we got to a point where we had data for almost all of the countries on the continent. The map was officially cluttered.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*bRseB4un_lvwps6lJI5_cg.png" /></figure><p>So, halfway through the year, with most African countries administering the Covax vaccines they received, we had to rethink our visualisation. How would we identify the countries that hadn’t received them?</p><p>We began using two maps: one with the bubbles of vaccines received and a second map highlighting the three countries that hadn’t yet received them. This worked for a few visualisations, but we knew we could simplify the process.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*lL0-26eYaV6RyRlJwTrgUQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DeWDzPJ5qp7z4qk9khAInA.png" /></figure><p>One of the basic principles of better visualisations is showing the data while also reducing the clutter. This is according to two of <a href="https://twitter.com/jschwabish">Jonathan Schwabish</a>’s five primary rules for better visualisations from his book <a href="https://policyviz.com/books/"><em>Better Visualisations: A guide for scholars, researchers and wonks</em></a>. (His remaining primary rules are integrating the graphs and text, avoiding the spaghetti chart and beginning with grey.)</p><p>Focussing on Schwabish’s first two rules, we needed to change our initial charts because there was just too much data on the graph.</p><p>On a separate map on <a href="https://theoutlier.co.za/covid-19/81619/36-african-countries-have-received-donations-from-the-us-since-july">US donations to Africa</a>, we adopted a different approach to mapping our data. There, we used a bar chart drawn in <a href="https://flourish.studio/">Flourish</a> with an .svg map of Africa colour-coded to match the bars.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*LgluMVjgTacu430-_ptSyw.png" /></figure><p>We liked how that one looked, so we adopted it for our weekly updates of Covax deliveries.</p><p>The result?</p><p>A more <a href="https://theoutlier.co.za/covid-19/81619/36-african-countries-have-received-donations-from-the-us-since-july">elegant chart</a> that could tell readers at a glance which countries had received the most Covax vaccines, the least and none at all in one graphic instead of two.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*h4ys-wJTFJzx6c4Nt-nKbQ.png" /></figure><p>So, next time your growing data starts to crowd out your map, have a rethink. You just might have to change your visualisation strategy with an eye on maximizing the data while minimising the clutter.</p><p><em>Media Hack Collective is launching in-person training in 2022. Watch out for more on data journalism training, tips and stories by </em><a href="https://newsletter.theoutlier.co.za/"><em>subscribing to our Outlier newsletter</em></a><em>.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cd82be6a97ba" width="1" height="1" alt=""><hr><p><a href="https://medium.com/media-hack-magazine/our-data-on-vaccine-deliveries-to-africa-was-too-much-for-our-map-to-handle-cd82be6a97ba">Our data on vaccine deliveries to Africa was too much for our map to handle.</a> was originally published in <a href="https://medium.com/media-hack-magazine">Media Hack</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What we learned from building a Covid dashboard for Africa]]></title>
            <link>https://medium.com/media-hack-magazine/what-we-learned-from-building-a-covid-dashboard-for-africa-bc75bdc6be06?source=rss-5a9f7c393039------2</link>
            <guid isPermaLink="false">https://medium.com/p/bc75bdc6be06</guid>
            <category><![CDATA[how-to]]></category>
            <category><![CDATA[vaccines]]></category>
            <category><![CDATA[africa]]></category>
            <category><![CDATA[data-collection]]></category>
            <category><![CDATA[covid19]]></category>
            <dc:creator><![CDATA[Gemma Ritchie]]></dc:creator>
            <pubDate>Tue, 22 Mar 2022 10:29:55 GMT</pubDate>
            <atom:updated>2022-03-22T11:12:05.901Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1018/1*tiO1eEGIkpV3B5U9tIET8Q.png" /></figure><p>South Africa reported its first Covid-19 case on <a href="https://www.nicd.ac.za/first-case-of-covid-19-coronavirus-reported-in-sa/">5 March 2020</a>. The government’s announcement was made in a media statement and posted on the internet as a PDF. That was pretty much how most of South Africa’s Covid data was made public — in PDFs and infographics.</p><p>For the most part, the daily updates were cumulative totals of confirmed infections, deaths, recoveries and tests done. Sometimes there were tables, but no time-series data — data recorded over consistent intervals of time — were made public.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/962/1*vUcktGhSzGPBwodxPG2gyA.png" /><figcaption>An example of the type of infographic released by the Department of Health to announce daily Covid-19 data</figcaption></figure><p>About a week after the first case was reported, when we realised the National Institute of Communicable Diseases (NICD) was issuing daily media statements of Covid stats, we decided to put that data into spreadsheets and visualise it.</p><p>So began what has become an ever-expanding data collection exercise which started with our South African <a href="https://mediahack.co.za/datastories/coronavirus/dashboard/">Coronavirus Dashboard</a> and gradually expanded to include vaccine delivery and use data from <a href="http://hosted.mediahack.co.za/evening-standard/vaccine-map/">all African countries</a>.</p><h3><strong>Data from the Global North</strong></h3><p>The <a href="https://coronavirus.jhu.edu/map.html">Johns Hopkins Covid-19 dashboard</a>, built by a collaboration of departments at Johns Hopkins University (JHU) in the United States, was the first resource that collected and made publicly available time-series data on Covid-19 in African countries.</p><p>After that, Oxford University’s <a href="https://ourworldindata.org/">Our World in Data </a>started to publish data, along with useful contextual information about why they chose to use one data source over another and how to interpret different data sets.</p><p>The data teams in big media companies in the US and Europe started to use data in visualisations that expanded the reach of Covid-19 data, giving their audiences access to public health data in <a href="https://www.washingtonpost.com/graphics/2020/world/corona-simulator">ways never seen before</a>.</p><p>These well-resourced organisations in the Global North established themselves as the sources of Covid-19 information and reporting on Africa. They set the news agenda and shaped how Africans were perceived outside and within the continent in the context of the global pandemic.</p><h3><strong>Why Africa needs to collect its own data</strong></h3><p>What we found, even with South Africa, was that the vaccine and vaccination data on the global dashboards was not as up-to-date as the data being published by African countries themselves or local WHO offices.</p><p>For instance, Our World in Data relied on data for the Democratic Republic of Congo supplied by the WHO’s Covid-19 dashboard from May — previously it had sourced its information from Kenyan daily The East African (over 1,800km away). However, for large portions of the year, the French regional WHO office (OMS) in the DRC gave Covid-19 data within its daily health reports. This information was invaluable as every number could be attributed to an official post.</p><p>So, for example, when we made widgets for three African countries to show daily vaccinations over time, we needed to collect our own, more up-to-date information put out by health authorities in African countries.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*cbnk3Kvt6KBB39U0xr67bQ.png" /><figcaption>Examples of some of the Covid-19 widgets we produced for a number of African countries</figcaption></figure><p>Generally, national health ministries would report confirmed Covid-19 cases and deaths, but collecting data on the types and number of vaccines African countries received was much harder.</p><p>With a bit of work, the Covax vaccine deliveries could be tracked. Unicef created a <a href="https://www.unicef.org/supply/covid-19-vaccine-market-dashboard">BI dashboard</a> that could be used as a reference point when the reporting of national authorities was sporadic.</p><h3><strong>How did we collect the data?</strong></h3><p>Collecting data for African countries was a labour-intensive, manual process. We started by compiling a list of 54 African countries with links to their health departments’ websites and social media accounts, particularly Facebook and Twitter. We also found links to World Health Organization (WHO) offices that posted vaccination data.</p><p>Some countries’ health departments had dashboards that they updated daily, such as Equatorial Guinea and Ghana, but they were few and far between.</p><p>Many put out their vaccination statistics as infographics on their social media accounts — like Kenya and Rwanda — or in press releases, such as Namibia and Côte d’Ivoire. Sometimes we had to watch videos of media conferences that were posted by news organisations, particularly to get information about vaccine deliveries.</p><p>Using the “<a href="https://twitter.com/i/lists/1369198335740018689">List</a>” function on Twitter we scraped and filtered the data. Initially, we scraped the data by running our Twitter list through an online tool called Workbench Pro but late last year Workbench shut down. After Workbench stopped, we had to find alternative ways to collect data from Twitter and Facebook updates from countries’ health ministries, including doing a lot of it manually.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*pMfbCBNldGery0NqtEUNAg.png" /><figcaption>Some of the sources we collect Covid-19 data from each day</figcaption></figure><p>Each morning and evening for more than a year, we have checked our list of sources for updates.</p><p>To help with the data collection, we also created Google email alerts for news articles written about vaccine deliveries.</p><h3><strong>Relying on trusty spreadsheets</strong></h3><p>We created two spreadsheets. In the first sheet we had the following variables: country, type of vaccine, date the vaccine was received, number of doses received, was the vaccine bought, donated or received through Covax. In the second sheet we had the variables: when did vaccinations start, how many vaccines had been administered, how many people were fully vaccinated which included the administration of the second doses of a two-dose vaccine and the one-dose Johnson &amp; Johnson.</p><p>Not all countries report all this information daily, so we have no idea how complete our data is. We do, however, note when our country’s data was last updated, so our reader can see how old the information is because sometimes there are long gaps between entries. It is particularly difficult to track the vaccines bought by countries directly from pharmaceutical companies and this information is often added later following announcements made by a country’s health minister or president.</p><p>But we have been able to put together a good enough data set that allows us to know which vaccines are being used in which countries, and we have a fairly good idea of the rate of vaccinations.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*zuwl1iNuISL2SN_0aaiaTQ.png" /><figcaption>African vaccine tracker created for the Evening Standard in London</figcaption></figure><p>We also use the <a href="https://africacdc.org/covid-19-vaccination/">Africa Centres for Disease Control and Prevention</a> and the <a href="https://covid19.who.int/">WHO</a> as reference points, just to doublecheck that we have the most up-to-date information. However, there are times when the data on those two dashboards appear to be behind. This is when it becomes critically important to audit this data against official sources such as the ministry of health or the country’s leadership which is typically released through the country’s official social media accounts.</p><p>We keep meticulous notes on where we source our data and include the links in our spreadsheets, so we can be sure that our data is the most reliable we can find.</p><p>As journalists, we also do our best to verify our sources, particularly when we are dealing with sources such as media organisations we are not familiar with in other countries.</p><p>We have a hierarchy of sources we use when we come across different numbers for the amounts of vaccines administered or delivered. First, we use a country’s official health department numbers. If those aren’t available, we use data from either the Africa CDC, the WHO or Unicef — whichever one posts about the vaccines administered or doses delivered.</p><h3><strong>The importance of making data accessible to the public</strong></h3><p>The impact of data journalism during the pandemic cannot be underestimated. Many people relied on media sources for reliable information with which they could make basic life decisions: How dangerous is it to go to the grocery store? Should I take my family on holiday this year? Is it safe to go back to the office?</p><p>Media companies like the <a href="https://www.ft.com/content/a2901ce8-5eb7-4633-b89c-cbdf5b386938">Financial Times</a> and <a href="https://www.nytimes.com/interactive/2021/world/covid-cases.html">The New York Times</a> produced exceptional stories and visualisations with Covid data.</p><p>But in many African countries, even academics, researchers and government decision-makers do not have easy access to data.</p><p>And it’s not just Covid-19 data, although the pandemic has brought the dearth of quality data to the forefront.</p><p>Catherine Kyobutungi from the African Population and Health Research Center told <a href="https://theconversation.com/wins-missteps-and-lessons-african-experts-reflect-on-two-years-of-covid-response-178904">The Conversation Africa</a> that public health communication in Africa had failed.</p><p>“The sudden avalanche of all sorts of information created an environment where complex concepts and knowledge that have traditionally been the preserve of academics and other experts became commonplace and hence open for gross misinterpretation. Public health communicators failed to adequately respond to this challenge,” she said.</p><p>Michelle Groome from South Africa’s National Institute for Communicable Diseases agreed that transparency and trust go hand in hand: “As scientists, we need to ensure that data is shared in a way that is easily understandable, that epidemiological terms are correctly interpreted, and that the rationale behind public health decisions is clarified. Lack of transparency and misinformation have led to lack of public trust in <a href="https://theconversation.com/what-south-africans-should-know-but-dont-about-governments-covid-19-vaccine-procurement-155756">government decisions</a> and non-acceptance of public health interventions.”</p><p>What we learned while collecting Covid-19 data for Africa over the two years of the pandemic can be applied to data on many other topics.</p><p>In South Africa, for example, time-series temperature data is not easy to access, while information on <a href="https://mediahack.co.za/2018/10/the-school-toilets-limpopo-forgot/">sanitation facilities in public schools</a>, the <a href="https://passmark.org.za/asbestos/">presence of asbestos in schools</a> and the number of renewable power plants in South Africa is difficult to collect. Often data sets are radically different, making it impossible to know which one is correct (or the most correct).</p><p>Collecting all this data, making it machine-readable and verifying it is a labour-intensive, time consuming and expensive process. But lessons learned from the Covid-19 pandemic show that for media companies, it can be a valuable resource for stories that can build trust among their audience and help provide resources to counter the disinformation and misinformation that proliferates on social media. — with additional reporting by Laura Grant</p><h4>For more on data journalism and data collection subscribe to our <a href="https://newsletter.theoutlier.co.za/">Outlier newsletter</a>.</h4><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bc75bdc6be06" width="1" height="1" alt=""><hr><p><a href="https://medium.com/media-hack-magazine/what-we-learned-from-building-a-covid-dashboard-for-africa-bc75bdc6be06">What we learned from building a Covid dashboard for Africa</a> was originally published in <a href="https://medium.com/media-hack-magazine">Media Hack</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>