Drop crazy yet little feasible ideas below

[sayantikabanik]

Idea 01:

• Quantifying psychological params and emotions (empathy, love, sad, anger etc)
• Dataset: Audio/video clips, English language, less than 10 secs. Twitter could be the source.

Idea 02:

• Impact of the pandemic on IT developers
• Dataset again thinking tweets

Idea 03:

• Potential suicide/ depression in teens
• Social media data

[vaibhav-code-py]

Idea 1:

• Predicting disruption in agricultural produce over the coming years due to pandemic.
• Dataset – https://data.gov.in/resources/district-wise-season-wise-crop-production-statistics-1997/api

Idea 2:

• Impact of Increased Packaged food sales on growing disease pattern (could take up specific disease like diabetes, cholesterol, cancer,etc).
• Dataset - swiggy/big basket/grofers api, Historical data on specific disease.

Idea 3:

• Predicting movement of currency USD/INR based on macro-economic events.
• Dataset - Historical data of currency, twitter data, Financial data.

Idea 4:

• Identifying pattern in various industries to predict employment opportunities.
• Dataset - Historical data on employment(data.gov.in), data from Naukri/linkedin api.

[sayantikabanik]

Discussion (narrowed to top 4 ideas) + @sahithi02 to add more

1) Quantifying psychological params and emotions (empathy, love, sad, anger etc)

Twitter text data, emoji based

2) Potential suicide/ depression in teens and divorce and domestic violence

Social media data, new articles (headlines) outcome - hard to define

• finance status
• mental health
• insurance cost increments/ loss of insurance

3) Impact of Increased Packaged food sales on growing disease pattern (could take up specific disease like diabetes, cholesterol, cancer,etc).

4) Identifying pattern in various industries to predict employment opportunities.

[sayantikabanik]

Idea 2:

• Impact of Increased Packaged food sales on growing disease pattern (could take up specific disease like diabetes, cholesterol, cancer,etc).
• Dataset - swiggy/big basket/grofers api, Historical data on a specific disease.

@vaibhav-code-py @sahithi02 Here are my thoughts after doing some googling

• didn't find anything directly matching with our idea
• hard to get any data from grofers, swiggy etc
• probably we need to think more about how to source the datasets and correlate the findings
  • Packaged food consumption data
  • Health data

[sahithi02]

Discussion (narrowed to top 4 ideas) + @sahithi02 to add more

1. Quantifying psychological params and emotions (empathy, love, sad, anger etc)

Twitter text data, emoji based

2. Potential suicide/ depression in teens and divorce and domestic violence

Social media data, new articles (headlines) outcome - hard to define

• finance status
• mental health
• insurance cost increments/ loss of insurance

3. Impact of Increased Packaged food sales on growing disease pattern (could take up specific disease like diabetes, cholesterol, cancer,etc).
4. Identifying pattern in various industries to predict employment opportunities.

5.Adding to the second idea of potential suicide /depression in teens . A facial detection system identifying a depressing person or to know if they are suffering from anxiety by changes from his/her previous photos along with their social media foot print.(the last article is about using instagram images)

Articles: -https://www.semanticscholar.org/paper/Extraction-of-Facial-Features-for-Depression-among-Venkataraman-Parameswaran/4b7772286158d08316f036f563b6d4ad26798545 -https://www.sciencedirect.com/science/article/abs/pii/S0165178113001017 -https://www.researchgate.net/post/Where_to_find_photo_dataset_of_depressed_people_faces -https://www.frontiersin.org/articles/10.3389/fnins.2021.609760/full -https://epjdatascience.springeropen.com/articles/10.1140/epjds/s13688-017-0110-z

[sayantikabanik]

final project

Identifying depression markers via social media and building an early-stage recommendation engine

• Resources
• https://www.kaggle.com/astraszab/facial-expression-dataset-image-folders-fer2013
• https://www.researchgate.net/post/Where_to_find_photo_dataset_of_depressed_people_faces
• https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6111060/

[sayantikabanik]

Link to proposal

https://isbhydmoh-my.sharepoint.com/:w:/g/personal/sayantika_banik_ampba2021w_isb_edu/Ebq5yBtX_8ZIjpTgMEIfM6IBdmL61n2K0eGWj1Kl3ou6pA?e=RUzgYX

[sayantikabanik]

keeping it open for reference