Journalistic Data Analysis#
User Story#
As a data journalist, I want to analyze and visualize parliamentary proceedings, decrees, consultations, political actors and votes in parliament in order to inform and/or entertain my readers.
Problem and Status Quo#
In the Swiss political system, important political decisions are made at the subnational (cantonal or municipal) level (cf. subsidiarity). However, the effort for data journalists to obtain subnational legislative data is disproportionately large (i.e. time-consuming and/or costly). Data is mostly not machine-readable, not harmonized and dispersed across 26 cantonal and 461 municipal parliaments. Thus, many news articles containing data analyses and visualizations focus on national data, missing an important part of the picture, and efforts of data journalists to obtain lower-level data are often one-off efforts and create duplicated, but not harmonized datasets, which is very inefficient.
Use Case#
flowchart TB
Parl1["Most Parliaments
(Data Providers)"] -- Web Crawling --> OPD["OpenParlData.ch
(Data Intermediary)"]
Parl2["Some Parliaments
(Data Providers)"] -- API --> OPD
Parties["3rd Party Data Providers"] -. API .-> OPD
OPD -- API --> User1@{ label: "Data Journalist 1
(Data User)" }
OPD -- API --> User2@{ label: "Data Journalist 2
(Data User)" }
User1 <-. share/collaborate .-> User2
User1 -. API (enriched data) .-> OPD
User2 -. API (enriched data) .-> OPD
style OPD stroke:#BBDEFB,fill:transparent
style Parties stroke-width:2px,stroke-dasharray: 2
OpenParlData.ch imports, cleans, harmonizes and openly publishes data from Swiss parliaments. It currently offers public data on political actors, parliamentary proceedings, decrees, consultations, votes and more from 78 parliaments. Using our API supplying harmonized legislative data, data journalists can analyze and/or visualize e.g. …
(topical) trends and political discourse in parliaments,
progress and success of bills e.g. relating to certain topics or by specific political actors at different federal levels,
politicians’ and parties’ actions (e.g. voting behavior), and thereby…
efficiently inform and/or entertain their readers and contribute to an informed public.
The harmonized structure of our data also enables data journalists to share cleaned and/or enriched data interoperably as well as topic models and data processing pipelines. This enables reproducing existing analyses and visualizations for different parliaments (i.e. regions).
Current scale#
The latter has already happened with a data journalist from Le Temps using our data to analyze voting behavior in the Grand Conseil of Geneva after having already done the same analysis for the National Council. Furthermore, a journalist from 20 Minuten already built an LLM for journalists using our API: https://huggingface.co/spaces/tomvaillant/cojournalist-data. It is currently in private beta.