- Goal of the project
- Future of the project
- Examples
- Useful links
- The goal of the project was to create a system for extraction of information from the ParlaMint and ParCzech corpora and storing them efficiently, allowing users to run queries on them.
- From these corpora, multiple properties of individual speakers can be observed (average words per speech, average length of speech, ...)
- The system should be able to extract the metadata about individual speakers (names, speeches, affiliations) and organisations (names, roles).
- System should provide an API for querying the data extracted from the corpora, which will be stored in a PostgreSQL database.
- It would be quite hard for a non-technical user to query the SQL database as they typically do not know the implementation details of the database, so a simpler way of querying should be designed and implemented into the API as well.
- The application, as it is now, is a console application with all servers running locally, all data stored locally.
- Further improvements, among other, could target this particular aspect.
- To be precise, the final goal is to create an application, which would allow users to conduct some statistical research upon the ParlaMint and ParCzech corpora, ideally in the form of web application.
- The big advantage of the ParaLint corpora is, that it has strictly defined format, and the corpra are available for many european languages allowing even cross-language statistic research and analysis.
- The goal of the first part of the project (current state) was to design and implement a backend for such web application.
- Additionaly, ParCzech corpus contains audio recordings of the speeches as well, which could bring many more interesting statistical insights, and which are not extracted by the application yet.
- So all-in-all improvement I would like to do:
- Design and implement convenient GUI for better and more enjoyable user experience.
- Create a frontend of the web application and link it to the backend.
- Run the server scripts on some actual server, not locally.
- Provide more means of visualization of the retrieved data (graphs / charts other than bar chart)
- Include the processing of audio as well, where it is present and possible.
- How many tokens has average speech for each gender?
+---------------------+------------+ | average_tokens | person.sex | +---------------------+------------+ | 84.0855333685910722 | M | | 35.2678488704064813 | F | +---------------------+------------+
M
- maleF
- female Data extracted from Latvian ParlaMint corpus. - How many speeches were given by members of each political party?
Data obtained from the ParCzech corpus.
+---------------------------------------------------------------+----------------+ | organisation.name | total_speeches | +---------------------------------------------------------------+----------------+ | Poslanecký klub ANO 2011 | 108813 | | Poslanecký klub České pirátské strany | 45663 | | Poslanecký klub České strany sociálně demokratické | 98991 | | Poslanecký klub KDU-ČSL | 48168 | | Poslanecký klub Komunistické strany Čech a Moravy | 82236 | | Poslanecký klub Občanské demokratické strany | 79395 | | Poslanecký klub Starostové a nezávislí | 12588 | | Poslanecký klub Svoboda a přímá demokracie | 29382 | | Poslanecký klub TOP 09 | 11862 | | Poslanecký klub TOP 09 a Starostové | 56895 | | Poslanecký klub Úsvit - Národní Koalice | 8268 | | ANO2011 | 15711 | | Česká pirátská strana | 39651 | | Česká strana sociálně demokratická | 68667 | | Komunistická strana Čech a Moravy | 76815 | | Křesťanská a demokratická unie - Československá strana lidová | 42672 | | Občanská demokratická strana | 59469 | | Starostove a nezavisli | 4356 | | Svoboda a prima demokracie - Tomio Okamura | 3483 | | TOP 09 | 55395 | | Úsvit přímé demokracie Tomia Okamury | 7578 | +---------------------------------------------------------------+----------------+
- Top 10 most talkative male speakers (most speeches given) Data obtained from the ParCzech corpus.