ANLY482 AY2017-18T2 Group 25 : Project Findings
| Interim | Final |
Data Preparation -> Data Cleaning -> Exploratory Data Analysis -> Next Steps To Take
Data Preparation: Received 4 different data sets
1. Reservation by users
2. Reservation mapping
3. Email campaign data
4. Email campaign user interaction
Data Cleaning: Methods Explored
- Used JMP Pro 13 to open all the data sets. However for email campaign user interaction data sets, even after it was spilt into monthly data, the file was still too big to be opened in JMP Pro, only able to open on Microsoft Notepad to take a look on the data.
- Our client told us that they used json and jupyter for their data and hence the team decided to explore using Pyton to clean our data.
- Team managed to clean the data in Pyton
Exploratory Data Analysis: Methods Used
- Softwares used were Jupyter and JMP Pro
- Exploratory Data Analysis was conducted to understand which variables were meaningful, which to be included in the analysis and that will allow us to achieve the objective
- Outliers were also found during this phase and possible causes for these outliers
- To exclude outliers from future analysis where applicable
Reservations Data:
- People tend to book in smaller groups of <10
- People tend to only reserve the table less than a day before the actual booking
Email Campaigns:
- Different types of emails have very varied open and click through rates
- Majority of customers receiving the emails do not open them
Next Steps To Take:
- For reservations data, to focus on the the key areas (e.g within a day what is the different recommendations)
- For reservations data, to map the actual restaurant names to the booking to look for trends
- For email interactions data, to map the time difference from receiving the email campaign to opening it
- To map both data sets to see conversion rate of customers
