🐦 Twitter dashboard

A Panel-based dashboard visualizing geotagged tweets with hvplot, Datashader and Echarts.

The dashboard includes:

• An heatmap showing the number of tweets;

• A bar plot showing the 5 most common languages within the current map extent;

• A wordcloud SVG showing the 10 most popular hashtags within the current map extent;

• A pie chart showing the overall sentiment (positive/negative) within the current map extent;

• Two line charts showing the number of tweets and unique users on a daily basis within the current map extent;

Deployed on

Hugging Face: https://huggingface.co/spaces/ivn888/Twitter-dashboard

Twitter data (link)

This dataset contains over 200k geotagged tweets in parquet format:

• Coverage: Historic Centre of Rome, Italy;

• EPSG: 3857;

• Year: 2018;

• Language: Multi;

• Source: This dataset was scraped by myself with snscrape;

Column	Description
tweet_date (index)	Tweet creation date
user_id	Unique identifier of the tweet author
user_location	User location information
tweet_id	Unique identifier of the tweet
tweet_text	Tweet content
tweet_hashtags	Comma separated list of the tweet hashtags
tweet_lang	Tweet language
tweet_sentiment	Sentiment Analysis (Multilingual model)
tweet_topic	Topic Classification (English, single-label model)
x	x-coordinate of the tweet
y	y-coordinate of the tweet

How to run this app on your local system

1. Git clone this repository:

git clone git@github.com:ivandorte/panel-geodashboard-twitter.git

cd panel-geodashboard-twitter

2. Install the required Python packages:

python -m pip install -r requirements.txt

3. Run the app

panel serve app.py --show

The dashboard will be available in your default web browser!!!

References

• HoloViz

• hvplot

• PyViz Topics Examples

• Awesome Panel

• TweetNLP

• Echarts

• Markdown Badges

Authors

• Ivan D'Ortenzio