When parliamentary monitoring organizations (PMOs), academicians and journalists visualize parliamentary votes, they usually display results of individual votes in some colorful or interactive way. But there is much more that can be seen if the votes of individual MPs are available.
One really cool visualization that we have developed at Kohovolit.eu, and which has been featured on this blog previously, displays patterns of voting using spatial proximity. It was developed by political scientist Keith Poole to study the US Congress in multiple ways. It has been used to research other parliaments, e.x. French or Latin American, the European Parliament or the Czech parliament (see an example below), as well as to conduct comparative studies.
This visualization can be either static or animated. In the former case, it covers a long period of time (e.g. one parliamentary term). In the latter case, the term is split into shorter periods of time and changes in MPs positions between these periods are animated.
The input of the visualization is a dataset of voting results. The output is a space where each MP is represented by a point. Distance between a pair of MPs shows how often these two MPs vote identically. If two MPs occupy the same spot in this space, they always vote together. The more distant the MPs are, the less often they vote together.
Positions of MPs on dimensions of this space do not have any a priori meaning – they depend on what the actual cleavages in the voting record are. Usually, one of the dimensions depicts the government-opposition split, but geographical, ethnic or religious conflicts between MPs and parties have been detected in various parliaments as well.
This visualization is priceless for parliamentary monitoring since it can help us devise theories about what is really going on in a parliament and displays parliamentary voting behavior in a way that is easy to understand and attractive to look at. It can be used to convey a number of messages, e.x. that parties cluster together on an underlying policy or that “rebels” or non-affiliated MPs switch their loyalties.
Apart from voting in national parliaments, the visualizations can also cover any group of people that votes frequently, e.g. United Nations General Assembly, local parliaments, national bank executive boards or even constitutional courts if judges’ dissent opinions are known.
Send us your data, we’ll send you a motion chart
KohoVolit.eu, a Czech and Slovak PMO, would like to offer the parliamentary monitoring community free visualizations. Just send us your vote data - for any parliament you like - and we’ll send you the visualization. You can link to it or embed it in a website or blog post. Journalists love it!
How should the data look? In order to create a colored and animated visualization, a CSV file with three simple tables is needed. First is a table where each row is an MP. The first column is a unique ID of the MP and the rest of the columns are votes. Votes must also have a unique ID recorded either in the first row or in the header of the table. See an example here.
The second table describes when votes were taken. The first column contains vote IDs and the second column contains the dates that votes took place in some recognizable format. See an example here.
The third table contains information on MPs’ memberships in parliamentary party groups. The columns are: 1) an MP’s ID, 2) a family name, 3) other names, 4) a full name of a parliamentary party group, 5) an abbreviation of the party group (ideally a widely recognized one), 6) a date when the membership started, 7) a date when the membership ended (the last column is left empty if the MP is still a member).
If an MP leaves a party group and enters another group, each membership is recorded into a separate row. For example, an MP that has switched party groups twice will have three rows in the table. Non-affiliation to party groups is treated as a membership in a “Non-Affiliated” party group. Therefore, if an MP is non-affiliated for a period of time, it needs to be recorded in a row. The table must include all MPs in the given period of time, including those who leave the parliament at some point. See an example here. Documentation for the tables is available here.
Send the tables to KohoVolit.eu and expect the visualization in return. Your data, which will be accessible from this Google doc, will also be uploaded to OpeningParliament.org and will remain available to anyone interested in reusing it. We strive to create an international archive of parliamentary data and we want to start with voting results since they’re very simpl,e but immensely interesting. A gallery of visualizations is available here and will also be made available at OpeningParliament.org.