Select Page

Twitter Analytics:

Talk-of-the-town in Edmonton

Application of Geo-tagged Tweets on Analyzing the Mental Health and Well-being of Residents: A Case Study on Edmonton 

Background: Mental Health and Wellness in Canada

Who is affected?

Mental illness indirectly affects all Canadians at some time through a family member, friend or colleague.
In any given year, 1 in 5 people in Canada will personally experience a mental health problem or illness.
Mental illness affects people of all ages, education, income levels, and cultures.
Approximately 8% of adults will experience major depression at some time in their lives.
About 1% of Canadians will experience bipolar disorder (or “manic depression”).

Source: Canadian Mental Health Association

Costs to society

The economic burden of mental illness in Canada is estimated at $51 billion per year. This includes health care costs, lost productivity, and reductions in health-related quality of life.
Individuals with a mental illness are much less likely to be employed.26 Unemployment rates are as high as 70% to 90% for people with the most severe mental illnesses.
The cost of a disability leave for a mental illness is about double the cost of a leave due to a physical illness.
In any given week, at least 500,000 employed Canadians are unable to work due to mental health problems. This includes:

  • approximately 355,000 disability cases due to mental and/or behavioral disorders.
  • approximately 175,000 full-time workers absent from work due to mental illness.

 

Adult Mental Health Disorder Prevalence in Canada

Major Depression

Bipolar Disorder

Schizophrenia

Anxiety

Source: Canadian Mental Health Association website: https://cmha.ca

Source: https://globalnews.ca/news/1766762/edmonton-is-the-saddest-city-on-twitter-study/
Source: https://www.cbc.ca/news/canada/edmonton/edmonton-the-saddest-canadian-city-on-twitter-1.2893814
Source: https://edmonton.ctvnews.ca/university-study-suggests-edmontonians-post-the-saddest-tweets-1.2181863

Edmonton: is it really the saddest city in Canada?

Check out the research project from the University of Ottawa published in 2014. It claimed that Edmonton is the saddest city in Canada!

The news grabbed the headlines of major newspapers and media outlets (see the screenshots above).

A group of researchers at UOttawa, led by Dr. Abdulmotaleb El Saddik, studied twitter data, hoping to increase the knowledge smart cities have about their citizens’ quality of life.

Edmonton was labeled as the saddest city (based on the research of tweets).

According to Dr. Abdulmotaleb’s research lab “The researchers used a four-dimensional vector analysis involving the four elements of pleasantness, arousal, dominance and unpredictability. Researchers also used emoticons and hashtags when collecting messages. In total, they compiled and categorized 47 different “emotion-word hashtags” (e.g., #angry, #happy, #disgusted) as well as their relationship to the other words in the tweet. The data comprises 132,181 tweets from December 4 to December 10, 2014”

Application of Machine Learning:

Classifying Health-related Tweets + Sentiment Analysis

What are we doing?

  •  Collecting Tweets across different time spans
  • Applying machine learning to sort relevant tweets
  • Analyze content to classify the emotions
  • Juxtapose data on urban infrastructure information
  • Visualize data through interactive map

How are we doing it?

  • Twitter data collection –  Grebe & Twitter API
  • Location prediction:
  • Using Support Vector Machines, Neural Network, Latent Dirichlet Allocation, Random Forest
  • Sentiment Analysis
  • Using multi-label emotion mining, VADER  lexicon

Why we are doing it?

To have a better understanding of the following questions:
  • How are YEG Tweets spatially distributed across neighborhoods?
  • How can we classify Tweets related to mental health?
  • How can we analyze the content of the Tweet?
  • What can we learn from Sentiment Analysis?
Disclaimer: This web application is a demo prepared for City of Edmonton’s HealthHack competition. Information might not be up-to-date. This is a working prototype. Content is based on datasets from Open Data Portal. No copyright infringement intended.

Team members:

Fahim Hassan (PhD Student, School of Public Health, University of Alberta). Also working in Ministry of Advanced Education, Government of Alberta.
Hamman Samuel (PhD Candidate, Department of Computer Science, University of Alberta). Also working in BioWare (video game developing company).
Mohib Khan (MSc, Engineering Management, University of Alberta, 2015). Also working in Ministry of Energy, Government of Alberta.
S A Rokib (MSc, Transportation Engineering, University of Alberta, 2015). Also working in Ministry of Transportation, Government of Alberta.