With the growing interests of young generation towards Bundestag (plus from our research, interviews, and study), we believe that there is a need for open data analysis, especially concerning what is happening within the Parliament. We hope that this product would help more people to care for, understand, and involved policy-making bodies, thus in turn becoming a more informed, engaged citizen.
What it does
VisGov is a chatbot that allows citizens to inquire information retrieved from the Parliament protocol in a simple manner, which tends to be a tedious and mundane reading process. Research survey we developed:
How we built it
For frontend, we use Angular.js, have it integrated with Dialogflow API. Supported by database analyzed by Natarual Language Toolkit(NLTK) using python and GENSIM(Topics Modelling for humans).
We used the open data from https://www.bundestag.de/service/opendata and https://code.holnburger.com/bundestag/auswertung.html#lange-der-reden
Challenges we ran into
We ran into challenges in integrating dialogflow with React.js framework. So out of time constraints, we switched to using angular framework, which works like a charm except we are still in the process of finding a solution to bind the properties of element to interactive elements of third-party libraries, such as d3.js. We also need a bigger database to train the intent in order to enhance the accuracy of bot's response.
Accomplishments that we're proud of
Being able to form a dedicated team of diverse background and expertise, which serve as our strength in coming up with this idea that can serve our city well. Being able to make a workable product within 30 hours.
What's next for VisGov
Potentially develop multilingual features that would reach a wider audience. Develop a more mature database that can enhance reliablity and credibility of response. This can be applied not just in Parliament, but even in ministries(if data is provided of course!) https://vimeo.com/user92339245/review/305079110/44c6266259