Column

Data Collection

1. The passive data I collected is the daily step. The data source is the Wechat Sports. Wechat Sports sends me my daily steps report everyday at around 10pm, and I take out my steps data from April 15 to May 12.

2. The administrative data I collected is the daily expenditure which is obtained through the Chase trasanction statement. I picked the transaction data from April 15 to May 12 and summarized all transactoins by day.

3. The active data I collected is the time when I went to bed. I actively recorded this data on the phone Notes App. I often recorded the time I went to bed when I turned off my bedside lamp and then closed my eyes.

Experiment

As I live off campus, I often take lunches to school several times a week. I want to know if I do not take any lunch to school, how it influences my life. So I set the experiment intervention as not taking lunch to school for all the experiment days.

My hypothesis is as follows:

1. If I do not take lunches to school, I’ll not spend time preparing for lunches in the evening. So I’ll go to bed early.

2. I need to walk to buy lunches if I do not take lunches, so there will be more daily steps.

3. As I need to buy lunches, my daily expenditures will be higher.

Column

Column

Column

Reflection on Experiment results

1. The pre-intervention mean of daily expenditure is 27.25214, and the post-intervention mean of daily expenditure is 30.98357. The postmean is higher than the premean, which is the same as my hypothesis. But after two-sample t-test, I got the p-value as 0.754, which means the two means are not statistically significant.

2. The pre-intervention mean of daily steps is daily expenture is 5825.571, and the post-intervention mean of daily steps is 5489.286. The postmean is less than the premean, which is opposite as my hypothesis. After reflection, I find out the reason behind the result. When I took lunches to school, I often had luch with my boyfriend in his department. So I often went from my department-Condon Hall to his department, which is around 15 minutes walking time. However, when I did not take lunches, I often grabed lunch near my department, so the average of daily steps is a little bit less before the intervention. But after running two-sample t-test, I get the p-value as 0.7072, which means that those two means are not statistically significant.

3. The average time I went to bed before intervention is 100 minutes later than 22:30 PM, which is 00:10 AM. But the avergae time I went to bed after the intervention is 75 minutes later than 22:30 PM, which is 23:45 PM. So the result is the same as my hypothesis. But the p-value of the two-sample t-test is 0.1423, which means the difference is not statistically significant.

Relection on the experiment process

First of all, I learnt about three different types of data in practical exercises. Then, I tracked all data continuously. I still think my intervention is meaningful to my life. Because the difference in Bed Time is important to me. If I’m in a super busy period, I can choose not to bring lunch and save time to go to bed early. But the trade-offs can be that I will spend more money during that period. But if I do a second time experiment, I need to just record dining costs rather than all transactions, which I believe can make the difference before and after intervention more significant. Because from the bar chart, I can see some unregular days befor and after the intervention, which is caused by other unregular costs. If I exclude those effects, I can get better experiment results.

Reflection on Class

This class opens the R world to me. I started to use R to tell stories about numbers. To me, the most helpful package to visulize data is the “ggplot” package. Before using R, I just used Excel to draw simple graphs to visulize data. After learning R, I can visulize data in a more complicated way and pick data as I like. But I feel like the most difficult visulization to me is drawing the map, which I still do not grasp clearly. Even though I’m still not good at coding yet, I learned the basic ways to learn R and practice.

About the final project, I think it is a meaningful way to practically exercise organizational experiments and coding. I need to review previous class codes and lectures and practice by myself. Besides, I think it is interesting to reflect on my own life and find out connections between different things. I hope I can learn more about how to choose data at the very begining of experiment design. With a better experiment design, I get better experiment results and analyze data more effectively.