NLP-on-Mental-Health-Issues-in-Twitter-Data

We analyse the conversations and attitudes about mental health in Twitter discourse through Natural Language Processing.

What We learnt!

Applying *Natural Language Processing on Twitter data appears to be an effective tool for analysis of mental health attitudes and can be a replacement or a complement for the traditional survey methods depending on the specifics of the research question.

Problem Description:

What mental health topics do people discuss on Twitter? Twitter data to analyse mental health issues.

Read the dataset

• In order to capture Twitter data, we had to follow the next steps:

• We needed a Twitter application and hence created a Twitter developer account.

• After registration, we grabbed our API keys and access tokens from Twitter: Consumer Key, Consumer Secret, Access Token and Access Token Secret.

• Install rtweet package in RStudio environment.

• Ran the script with the API keys and access tokens as input parameters.

• The hashtags considered for this analysis are as following:

#mentalhealth , #depression , #worldmentalhealthday , #WMHD, #nostigma , #nostigmas , #eatingdisorders, #suicide, #ptsd, #mentalhealthawareness, #mentalillness, #stopsuicide, #IAmStigmaFree, #suicideprevention, #MH, #addiction, #bipolar, #stigma

• We were able to download almost 18K records on a single try.

Findings , Proposed Solution and Scope of Imorovement:

Findings

• On an average, almost 30 tweets are submitted every single second with hashtags related to mental health issues.
• We found the most common hashtag of them all is related to Suicide and Depression.
• Common Negative sentiments for such discussion were ‘miserable’, ‘desperate’, ‘distress’, ‘rape’, ‘pain’ etc.
• Most tweets came from the areas of United States, United Kingdom and Canada.

Proposed Solution

• Identify these tweets and provide them with support hotlines numbers immediately.

Scope of improvement

• Identifying users who are constantly posting about these negative sentiments and provide them with different help options like anonymous groups, help and support group information, doctors/therapist information privately in their emails. * This might encourage the user to seek the help that they might require.