How to visualize polls & results of the U.S. elections with Datawrapper

Lisa Charlotte Muth

Graphics reporters and journalists have been busy this year with reporting on the coronavirus pandemic – and now the U.S. elections are around the corner, too. On Tuesday, 3rd of November (and days and weeks before with lots of letters), U.S. citizens will decide on all House seats, 35 Senate seats, and if they want to call the Republican Donald Trump or the Democrat Joe Biden their next president.

It’s an important election. We expect you might want to cover at least the Presidential part of it – and that’s what we’ll focus on in this blog post –, even if you’re not from the U.S., and even if you don’t have the team or the capacity to create a big election dashboard.

Here are some ideas on which data to cover and how to do so.

A word ahead: We won’t offer any live-updating charts on election night. However, the friendly people from FiveThirtyEight put their election forecasts on Github. We used them in some of the following charts, maps, and tables. They will update once a day. Feel free to re-use all of the following visualizations, no matter if they come with live-updating FiveThirtyEight data or test data: Hover over the visualization of your choice, click on “Edit this chart”, change the data if necessary, and then publish & embed it.

Which states are undecided

While most people in many states have been voting either Democrat or Republican for years – and polls show that they plan to do the same this year –, “swing” or “battleground” states will decide the election. Most newsrooms focus on them and how little we know (so far) about who will win in these states. To do the same, you can show a map like the following:

Here we decided to use a symbol map instead of a choropleth map. After all, electoral votes will decide the election, not land area. When mapping election polls or results on a state level, consider using symbols on top of the state’s boundaries as we did. The symbol sizes represent the number of electors for this state.

In addition to a map, consider showing a table with detailed information. A table has the advantage to be sortable. Here we sorted it by the difference in the estimated votes for Trump and Biden, to show the swing states first:

Electoral votes

The U.S. Presidential election will be decided with electoral votes. So many newsrooms show a version of the following line chart:

You can give your readers a concise idea of how many electoral votes (are estimated to) go to which candidate with a stacked bar chart. You will find it in many election dashboards at the very top:

<b style="text-align:center; display:block; font-style:normal; margin-left:-1%; margin-top:-50px; font-size:15px;">⎹<br>⎹<br>270 votes are needed<br>to win the election</b><br><br>Electoral votes as forecasted by FiveThirtyEight. Darker colors mean more likely votes. Grey votes are from uncertain states. 270 electoral votes are needed to win the election. 

You can also split up the bars to show a plainer comparison that also works well for small screens:

When results will come in

Because of the coronavirus pandemic, lots of voters vote by mail. It will take days (maybe weeks?) for all the votes to be counted so that the election coverage will drag on. Consider letting your readers know when they can expect results and which states have already called a winner.

You can do the same in map tooltips:

The days and weeks after the election: Analysis

After the election comes the time where everyone analyzes what just happened. Consider the following chart types to do so:

• A stacked bar chart to show voter demographics (e.g., voters by age, religion, education, ethnicity)
• State maps to show the turnout
• County maps or scatterplots to show which counties shifted from voting Republican to voting Democrat and the other way round
• Arrow charts to show which states shifted the most

Here's part of what the stacked bar chart showing voter demographics could look like:

A word about maps

Maps are essential for election coverage. If you're a U.S. organization, your readers will want to know the results from their home county. That's why Datawrapper lets you create county maps for each U.S. state:

There are also lots of U.S. maps to choose from in Datawrapper. You've already met two in this article: the U.S. tilemap sized by electoral votes and the U.S. state map. Here are a few more:

• U.S. hexagons map, where each state has the same size. That's a good choice if the number of states is essential (e.g., how many states allow which kind of postal voting).
• U.S. congressional district map to report on the election results for the House of Representatives or the win probability before the election
• U.S. county map to show how individual counties voted (e.g., for the Senate), or how that differs to the 2016 election.

In total, you'll find more than 2000 maps in Datawrapper. You can see the whole collection in our map selector (no need to sign up). If you don't find what you're looking for, get in touch with us at support@datawrapper.de or upload your own GeoJSON.

Inspiration

To get more inspiration for how you could visualize election data, visit the dashboards that newsrooms created for the 2016 and 2018 elections: Guardian, BBC, Washington Post, Financial Times, Bloomberg, Politico, New York Times 2016, New York Times 2018.

You can also have a look at the poll trackers that some newsrooms have created for the 2020 election: the Economist, New York Times, Financial Times, BBC, Guardian, Washington Post, FiveThirtyEight.

We hope that these visualizations provide a good starting point for your election visualizations. To use any of the ones above, hover over them and click on "Edit this chart". This will lead you right to step 3: Visualize. From there, you can either change the data (source) and the title, labels, colors, etc. – or you can directly publish your visualization to get the embed code. Any questions? Let us know at support@datawrapper.de.

---

Lisa Charlotte Muth

(she/her, @lisacmuth, @lisacmuth@vis.social) is Datawrapper’s head of communications. She writes about best practices in data visualization and thinks of new ways to excite you about charts and maps. Lisa lives in Berlin.

Liked this article? Maybe your friends will too:

Twitter Facebook