Obesity is a serious health concern not just in the US but across the globe. Obesity is also associated with increased comorbidities leading to increased healthcare expenditure by the patients and also applies a significant burden on the hospital resources. Though we are gradually recovering from the COVID situation, the problem isn't solved yet. It is only part of the solution. We feel being in lockdown (with little/no exercise) may lead us towards the path of fighting another pandemic potentially leading to a burden on public health resources and individual income. Therefore, our topic fit better under this challenge "Integrated Assessment" which deals with the socio-economic impact of COVID-19.
Motivation
It was our personal experience during lockdown which lead us to take up this project on Obesity. Lack of physical activity and staying at home for more than 2 months made us realize how mentally tiring it can be. Hence, we started off with this project in an effort to investigate the role of sedentary behavior (lack of physical activity) during lockdown in making us obese. In addition, Looking at the challenges that SpaceApps had, we felt "Integrated Assessment" is the right challenge category for our problem. So, it was our problem question which led to us to sign-up under this challenge
Approach
Our interest was to study specifically the influence of physical activity on the Obesity rate in the US.
1) We initially extracted the annual average time use data of Americans (15 years and older) from 2008-2018 based on weekdays, weekends and gender (data provided under the "Resources" tab of the challenge) on activities listed below
2) As we couldn't find a nationally representative sample to conduct our experiment, we made use of Census data from 2016 for the USA to simulate a sample population of 10000 subjects. We used the Microsimulation modeling approach to generate sample data.
3) Demographic characteristics of the sample selected are provided in the PPT (link can be found below)
4) The simulation was done using Python Jupyter Notebook (link provided in the appropriate section).
5) We created 3 python classes such as
a) Population - Generates a sample population of 10000 subjects,
b) Calories - Calculates calory intake, fat burn to
c) Activities - Computes the probability a person is engaged in each activity.
6) Later, we ran what-if scenarios to know the obesity rate for different values of "average hours" spent under the "Exercise/Physical activity" section.
7) We verified our simulation with two years of data (2017,2018). Through this step, we were able to see that the obesity rate for simulated data wasn't very much different from the original data.
7) In Through our multiple experiments, we were able to see that there is a negative relationship between physical activity and obesity rate.
Resources used
To solve this problem, we used the resources provided by SpaceApps under the "Resources" section along with few more open data sources.
1) US Bureau of Labor Statistics - We used this data source to get data on Time spent by Americans on different activities like Eating, Sleeping, etc for the past 10 years (2008-2018)
2) https://www.cdc.gov/obesity/index.html
3) https://ourworldindata.org/obesity
4) https://www.cdc.gov/nchs/data/factsheets/factsheet_nhanes.htm
5) https://www.cdc.gov/nchs/data/factsheets/factsheet_nhanes.htm
6) https://www.medicalnewstoday.com/articles/245588#recommended-intake
7) https://www.medicalnewstoday.com/articles/319731
8) https://www.pbrc.edu/research-and-faculty/calculators/weight-loss-predictor/about/
https://docs.google.com/presentation/d/1WtIiHqjwYOdXgx3JvvFuDkkF8hJEUR80fX5LBQqn-2w/edit?usp=sharing
1) https://www.bls.gov/tus/tusa_2tabs.htm (Provided as a resource by SpaceApps)
2) https://www.cdc.gov/obesity/index.html
3) https://ourworldindata.org/obesity
4) https://www.cdc.gov/nchs/data/factsheets/factsheet_nhanes.htm
5) https://www.cdc.gov/nchs/data/factsheets/factsheet_nhanes.htm
6) https://www.medicalnewstoday.com/articles/245588#recommended-intake
7) https://www.medicalnewstoday.com/articles/319731
8) https://www.pbrc.edu/research-and-faculty/calculators/weight-loss-predictor/about/