Sentimental

Social Media Sentiment Analysis.

Because social media streams can need human response

About

Combination of

Twitter Streams
Sentiment Analysis
PagerDuty

Potential of adding additional Social Media Streams, Tweaking Sensitivity will come in the future.

Deployment

Preferred deployment via Docker or ECS.

Clone the repository with

git clone https://github.com/justynroberts/sentimental.git .

To get the latest repo You will need Docker installed also.

Build & Run

cd sentimental

To clean any previous builds

docker rm -f sentiment

The file example build.sh can be used to build and run a local docker image cd to src directory

docker build -t sentiment .

Docker Parameters

To run the container, you will need to execute the parameter with a set of configuration items, with an example below:

docker run -e MONITOR="@twitterhandle" -e PD_ROUTING_KEY="PagerDuty Routing Key" -e TW_CONSUMER_KEY="Twitter Consumer Key" -e TW_CONSUMER_SECRET="Twitter Consumer Secret" -e TW_ACCESS_TOKEN="Twitter Access Token" -e TW_ACCESS_SECRET="Twitter Access Token" -e TW_BEARER_TOKEN="Twitter Bearer Token" -e TW_TIMEFRAME="100" -d --name sentiment sentiment

Parameters Broken down as:

MONITOR="Hashtag or Twitter Handle eg @Justynroberts"

PD_ROUTING_KEY="Enter your PagerDuty Service Routing Key"

TW_CONSUMER_KEY="Twitter Consumer Key"

TW_CONSUMER_SECRET="Twitter Consumer Secret"

TW_ACCESS_TOKEN="Twitter Access Token"

TW_ACCESS_SECRET="Twitter Access Secret"

TW_BEARER_TOKEN="Twitter Bearer Token"

TW_TIMEFRAME="100" # Time window in seconds for repeat tweet before notification

Logs

To grab inflight logs

docker logs sentiment -f

Twitter

You WILL need a set of API tokens/secrets from developer.twitter.com

Caveats

This isnt particularly error robust, or configurable.. Currently Proof Of Concept. But it works.