An overview of the data
Data used in this project include: one-card consumption data, campus wifi data and meteorological data of Minhang District of Shanghai. Specifically, it includes:
- Merchant information: a total of 134 observations, including 32 merchant systems and 85 sub-merchants;
- User information: a total of 30861 observations, including 30861 one-card accounts and 30812 student numbers, that is, 49 students have two one-card accounts;
- Transaction records: a total of 7915289 transaction records, the time span from 2014-09-01 to 2015-01-31;
- Meteorological Records: 26,660 meteorological records, from August 15, 2014 to March 25, 2015;
- Campus network records: 1,2736,408 campus network records in total, ranging from 2014-09-01 to 2015-01-31.
My work
My work mainly focuses on one-card consumption data.
The first is the summary statistics of the data, including user group distribution, business structure, historical rainfall and campus network records.
From total consumption category, then the mean consumption category number, category, such as total amount of total consumption of the boys and girls consumption point of statistical records of canteen, and the third catering students with restaurants, for example, to analyze its historical repast amount change trend, rain and sunshine repast situation compared, the different times of day dining content such as density, To study the relationship between the amount of dining hall and rainfall, the number of active campus network and other factors.
Finally, it analyzes the consumption distribution of different types of users in various consumption categories and merchants, including total consumption and consumption times, and establishes two models, horizontal mode and vertical mode, to analyze the dining patterns of different types of users.
The final report is deployed here and the code is hosted on my Github.
Work to upgrade
Later, I realized that I should not only produce a data analysis report, but also hope to present a complete application-level system. The data source is no longer just one-card consumption data, but a comprehensive 360-degree analysis of one-card consumption and campus network wifi data.
After a month of hard work, my product, Elite, a real-time monitoring and analysis statistical system with the vision of building a smart campus data ecosystem, has gradually taken shape.
Elite means “Elite”, that is, we aim to extract the best part of the data to users, while “E” stands for digital and information technology, “Lite” is similar to “life”, that is, data life in smart campus.
In general, the functions of Elite include the following aspects: campus real-time monitoring, catering analysis, academic management, teaching statistics, personal consultant, Elite assistant, information sharing, etc.
- Real-time monitoring: including real-time crowd monitoring, crowd migration trend and dining population prediction, etc.
- Catering analysis: including the real-time statistics of the total amount of dining and dining times of the major merchants today, the calculation of the total amount of dining and the historical daily total, the composition of dining crowd and the analysis of the proportion of the major canteens;
- Academic management: including the summary of male and female students’ online keywords, the analysis of students’ attention to different keywords in different grades, the statistics of wifi traffic distribution of students in different scenes and at different times, the distribution of students’ one-card consumption and wifi network, real-time monitoring and prediction of bathroom crowd traffic;
- Teaching statistics: including campus equipment, research equipment, teaching equipment statistics, research projects, research funds, research works, published papers statistics, hard indicators of each campus, and the number of students of all types statistics;
- Personal consultant: obtain the student’s one-card consumption and campus wifi data according to the student id after anonymous processing, and display personalized statistical results from the perspective of diligence coefficient, dining coefficient, sleep coefficient, cleanliness coefficient, Engel coefficient, personal evaluation, personal concern, user group classification, etc.
- Elite Assistant: provides personalized recommendations and suggestions for users, including life early warning, learning early warning, activity recommendation, course recommendation, dynamic record, etc.
- Information sharing: including “activities”, “comments”, “competition”, “second-hand”, “rental”, “internship”, “school bus” seven modules, gathering dynamic information in all aspects of the campus.
As for the positioning of Elite, I think it is the connector between data source and sink in the smart campus. To be more specific, in the smart campus, an ecosystem of continuous data circulation, students and school authorities are data producers, students, school authorities and merchants are data consumers, while Elite is data decomsolver. Elite implements data processing and resource integration and reuse, enabling all actors in the ecosystem to help each other and win.
I used Prezi to make a product introduction and share. Elite is deployed here, and the code is hosted on my Github, where you can find prezi for product introductions and sharing.
conclusion
After this attempt, I have mastered the basic use of R and the process of data analysis, and strengthened my ability of data visualization. The results of the work gradually improved from simple statistical analysis at the beginning, and eventually grew into a perfect and powerful intelligent campus data analysis system. The workload was huge, which was a good development experience.