SLEEP PROJECT
"Project Description
In this project, you will think of a topic and problem, collect experimental data, complete an end-to-end analysis and present the results, all by yourself.
First, choose a topic of interest to you and understand what research has already been done in that area. What are some interesting questions that remain? Can you turn those questions into a product (i.e. can you extract value out of answering those questions)?
You will then collect some data you think could help answer those questions. Choose your main source of data wisely, since in this project you have a restriction that tries to emulate a common corporate setting: you won't have access to a census of the universe of your choice. You must collect the data yourself in such a way that the universe of datapoints available to you is limted. For example, you may be limited by time (e.g. watching and categorizing Youtube videos or Instagram pictures), by cost (e.g. querying Google Maps for public transport routes via the gui, without paying for the API access) or by access (e.g. surveying people on their preferences). In the end, you should aim at collecting between 30 and 100 observations (rows) and between 5 and 10 features (columns) per observation.
Once you have your data, complete an analysis that answers your original question and/or related ancilliary questions. Please make sure that the main observations you make hold to scientific scrutiny at some level of significance. You can and should supplement your analysis with visual intuition and highlights of hypothesis that the data seems to support, even if you are not necessarily able to hold those insights to the same level of scrutiny as your main question.
You can enrich your limited dataset with information from richer sources that you can obtain trough any means you've learned before (e.g. you may web scrape the weights of car models if that is one of your observations).
Like in the previous project, package your results with a product or service mindset. You will present your findings in a presentation (possibly supported by an interactive visualization) where you should evidence principles of dashboarding and storytelling.
Project Goals Research, collect and analyse data on a topic of interest to you. Feel free to use additional data to enrich your dataset, maybe using an API or web scraping. Apply the statistical techniques we have learned, along with techniques from EDA. Create useful and easily-interpretable plots. Prepare a presentation keeping in mind the finer points of storytelling. Communicate the results of your analysis clearly, accurately and engagingly."
In this project, I chose to analyze a Sleep Health dataset, with information such as BMI, Age, Gender, Occupation, Sleep Duration, Quality of Sleep, Physical Activity Level, Stress Level, Blood Pressure, Heart Rate, Daily Steps and Sleep Disorder.
Based on the information the dataset provided me, I tested the following hypothesis:
- Do people with higher BMI (obese and overweight) have more sleep disorders (Sleep Apnea/Insomnia) than people with normal BMI?" H0: There is no significant correlation of sleep disorders based on BMI category. H1: BMI category affects sleep disorders.
Tested the correlation of Sleep Disorders with BMI using Chi2 contingency table
- "Do people with age > 40 have better quality sleep?" H0: There is no significant association between age and sleep quality. H1: Age is associated with sleep quality.
Tested the correlation of Quality of Sleep and Age using Two-Sample T-Test
- "Is a higher physical activity level associated with higher sleep quality?" H0: There is no significant association between Physical Activity Level and sleep quality. H1: Physical Activity Level is associated with sleep quality.
Tested the relationship of Quality of Sleep and Physical Activity Level using Spearman's correlation
- "Is sleeping over 6 hours per night associated with lower stress levels?" H0: There is no significant correlation between stress level and sleep duration. H1: Stress level is negatively correlated with sleep duration.
Tested the relationship of Sleep Duration and Stress Levels using Pearson Correlation
After testing all of the hypothesis and having the results of the correlation relationships, I made a Canva presentation displaying my findings.
Sleep Health Project Presentation - https://www.canva.com/design/DAFrnTWYaU4/7gpOLe1ZcCZan_7hn2Xd1g/edit