search

Sentiments-of-Bundestag / graphenauswertung

0 stars 0 forks source link

readme

graphenauswertung

Prerequisites

Install the requirements

pip install -r requirements.txt

Running the App

python app.py

App runs on http://localhost:5000.

Run a local Neo4j database

Prerequisites

Docker

Start the Mock DB

docker-compose up

Visit the browser

http://localhost:7474/browser

username: neo4j password: graphenauswertung

Empty the database

MATCH (n) OPTIONAL MATCH (n)-[r]-() DELETE n,r

Endpoints

Method: GET for all

/sessions

returns all available parliament sessions. The sessionId can be used to filter other endpoints (see below).

Sample Data

[
  {
    "sessionId": 138, 
    "legislativePeriod": 19, 
    "endDateTime": "2019-12-20T15:39:00+00:00", 
    "startDateTime": "2019-12-20T09:00:00+00:00"
  }
]

/factions

returns all factions from group5 database

Sample Data

[
  {
    "name": "CDU/CSU", 
    "size": 246,
    "factionId": "F000",
    "sessionIds": [ 46, 100]
  }, 
  {
    "name": "SPD", 
    "size": 152,
    "factionId": "F001",
    "sessionIds": [ 46 ]
  }
]

/factions/graph

returns the faction graph based on the group5 database. Messages are aggregated so that there is only one entry for a unique (directed) relationship between sender and recipient. The count is the total number of messages that were aggregated to the entry. The sentiment is the sum of the sentiments of all messages that were aggregated to the entry.

Query Parameters

Name Description Allowed Values
filter Limits the messages to only those with the given type of sentiment POSITIVE, NEGATIVE
session Limits the messages to only those from the given session any sessionId from /sessions
excludeApplause Filters out all messages which are flagged as applause true

Sample Data

{
  "factions": [
    {
      "name": "CDU/CSU",
      "size": 246,
      "factionId": "F000"
    },
    {
      "name": "SPD",
      "size": 152,
      "factionId": "F001"
    }
  ],
  "messages": [
    {
      "recipient": "F001",
      "sender": "F000",
      "sentiment": -1.399999976158142,
      "count": 10,
      "sessionIds": [46, 100]
    },
    {
      "recipient": "F000",
      "sender": "F001",
      "sentiment": 0.1,
      "count": 2,
      "sessionIds": [46, 100]
    }
  ]
}

/factions/ranked

returns the page-ranked factions from group5 database

Query Parameters

Name Description Allowed Values
filter Limits the messages used for page rank calculation to only those with the given type of sentiment POSITIVE, NEGATIVE
session Limits the messages used for page rank calculation to only those from the given session any sessionId from /sessions
reverse Uses the ''Reverse PageRank'' for page rank calculation true
excludeApplause Filters out all messages which are flagged as applause true

Sample Data

[
  {
    "name": "CDU/CSU", 
    "size": 246,
    "factionId": "F000",
    "rank": "0.7"
  }, 
  {
    "name": "SPD", 
    "size": 152,
    "factionId": "F001",
    "rank": "0.3"
  }
]

/factions/proportions

returns speech proportions of each faction in percentage, regarding all sessions.

Name Description Allowed Values
session Limits the messages used for proportion calculation to only those from the given session any sessionId from /sessions
excludeApplause Filters out all messages which are flagged as applause true

Sample Data

[
  {
    "factionId": "F000",
    "name": "CDU/CSU",
    "proportion": 24.664010828634297,
    "size": 246
  },
  {
    "factionId": "F001",
    "name": "SPD",
    "proportion": 20.437122713830245,
    "size": 152
  }
]

factions/sentiment/key_figures

returns the calculated key figures of sentiments between all factions

Name Description Allowed Values
session Limits the messages used for key figures calculation to only those from the given session any sessionId from /sessions
excludeApplause Filters out all messages which are flagged as applause true

Sample Data

{
  "highest_sentiment": 16.04166666790843, 
  "lowest_sentiment": 0.5, 
  "sentiment_lower_quartile": 1.0, 
  "sentiment_median": 2.5903602056205273, 
  "sentiment_upper_quartile": 5.3949188850820065
}

/persons

returns all persons from group4 database

Query Parameters

Name Description Allowed Values
filter Limits the persons to only those connected to messages with the given type of sentiment POSITIVE, NEGATIVE
session Limits the persons to only those connected to messages in the given session any sessionId from /sessions
person Limits the persons to only the given person any speakerId from /persons

Sample Data

[
    {
      "faction": "CDU/CSU",
      "factionId": "F000", 
      "name": "Frank Heinrich", 
      "role": "Platzhalter", 
      "speakerId": "MDB-24aa7763-e95d-4d1d-834c-de3cae2406d7",
      "sessionIds": [ 46, 100 ]
    }, 
    {
      "faction": "SPD", 
      "factionId": "F001",
      "name": "Frank Schwabe", 
      "role": "Platzhalter", 
      "speakerId": "MDB-c0f339ee-9db1-411d-ad2f-0357e98bf112",
      "sessionIds": [ 46 ]
    }
]

persons/messages

returns all messages from group4 database. The messages are not aggregated in any way.

Query Parameters

Name Description Allowed Values
filter Limits the messages to only those with the given type of sentiment POSITIVE, NEGATIVE
session Limits the messages to only those from the given session any sessionId from /sessions
person Limits the messages to only those from the given person any speakerId from /persons
excludeApplause Filters out all messages which are flagged as applause true

Sample Data

[
    {     
      "recipient": "MDB-c0f339ee-9db1-411d-ad2f-0357e98bf112", 
      "sender": "MDB-24aa7763-e95d-4d1d-834c-de3cae2406d7", 
      "sentiment": 0.2,
      "sessionId": 46
    },
    {
      "recipient": "MDB-24aa7763-e95d-4d1d-834c-de3cae2406d7", 
      "sender": "MDB-c0f339ee-9db1-411d-ad2f-0357e98bf112",
      "sentiment": 0.7, 
      "sessionId": 46
    }
  ]

persons/graph

returns the person graph based on the group4 database. Messages are aggregated so that there is only one entry for a unique (directed) relationship between sender and recipient. The count is the total number of messages that were aggregated to the entry. The sentiment is the sum of the sentiments of all messages that were aggregated to the entry.

Query Parameters

Name Description Allowed Values
filter Limits the messages to only those with the given type of sentiment and the persons to only those connected to messages with the given type of sentiment POSITIVE, NEGATIVE
session Limits the messages to only those from the given session the persons to only those connected to messages in the given session any sessionId from /sessions
person Limits the messages to only those from the given person and the persons information to only those connected to the given person any speakerId from /persons
excludeApplause Filters out all messages which are flagged as applause true

Sample Data

{
  "persons": [
    {
      "faction": "CDU/CSU", 
      "factionId": "F000",
      "name": "Frank Heinrich", 
      "role": "Platzhalter", 
      "speakerId": "MDB-24aa7763-e95d-4d1d-834c-de3cae2406d7",
      "sessionIds": [ 46, 100 ]
    }, 
    {
      "faction": "SPD",  
      "factionId": "F001",
      "name": "Frank Schwabe", 
      "role": "Platzhalter", 
      "speakerId": "MDB-c0f339ee-9db1-411d-ad2f-0357e98bf112",
      "sessionIds": [ 46 ]
    }
  ],
  "messages": [
    {
      "count": 2, 
      "recipient": "MDB-c0f339ee-9db1-411d-ad2f-0357e98bf112", 
      "sender": "MDB-24aa7763-e95d-4d1d-834c-de3cae2406d7", 
      "sentiment": -0.2,
      "sessionIds": [ 46, 100 ]
    },
    {
      "count": 4, 
      "recipient": "MDB-24aa7763-e95d-4d1d-834c-de3cae2406d7", 
      "sender": "MDB-c0f339ee-9db1-411d-ad2f-0357e98bf112", 
      "sentiment": 0.1,
      "sessionIds": [ 46 ]
    }
  ]
}

persons/ranked

returns the page-ranked persons from group4 database

Query Parameters

Name Description Allowed Values
filter Limits the messages used for page rank calculation to only those with the given type of sentiment POSITIVE, NEGATIVE
session Limits the messages used for page rank calculation to only those from the given session any sessionId from /sessions
reverse Uses the ''Reverse PageRank'' for page rank calculation true
excludeApplause Filters out all messages which are flagged as applause true

Sample Data

[
  {
    "faction": "DIE LINKE",
    "factionId": "F003", 
    "name": "Caren Lay", 
    "rank": 0.029715244473085708, 
    "role": "Platzhalter",
    "sessionIds": [ 46, 100 ],
    "speakerId": "MDB-c3f825cc-9b63-4241-85f9-df425f0c6486"
  }, 
  {
    "faction": "CDU/CSU",    
    "factionId": "F000",
    "name": "Wolfgang Schäuble", 
    "rank": 0.029176551464191278, 
    "role": "Platzhalter",   
    "sessionIds": [ 46   ], 
    "speakerId": "MDB-fd366231-9c25-416b-8604-e934d956e177"
  }
]

persons/sentiment/key_figures

returns the calculated key figures of sentiments between all persons

Query Parameters

Name Description Allowed Values
session Limits the messages used for key figures calculation to only those from the given session any sessionId from /sessions
excludeApplause Filters out all messages which are flagged as applause true

Sample Data

{
  "highest_sentiment": 0.6166666666666667,
  "lowest_sentiment": -1,
  "sentiment_lower_quartile": 0.07142857142857142,
  "sentiment_median": 0.5,
  "sentiment_upper_quartile": 0.5
}