On December 21-25, 2018, CDA Data Analysis Institute conducted a learning activity on the theme of "Data Science Application" at Golden Arch (China) Co., Ltd. The number of applicants in all departments of the Shanghai headquarters exceeded expectations, and a total of 50 employees participated in the study. Teachers and relevant colleagues in the data analysis department actively communicated, and this activity was a complete success.
Study outline :
The first stage : Python programming and data collation essentials
1. Data structure (list, dictionary, tuple, etc.)
2. Loop structure (traversal of text), use function
3. Arrays, numpy, pandas, etc.
Case:
Customer dining coupon collection problem, customer self-avoidance random walking problem, customer normal distribution function problem
Count the number of characters in takeaway orders, the classification and summary of rental prices, the cleaning and collation of football player data, etc.
The second stage : sampling distribution, hypothesis testing and logistic regression model
1. Sampling distribution, random process and random simulation, parameter statistics and hypothesis testing of mean ratio
2. Processing of categorical variables, model parameter estimation (mastering maximum likelihood estimation)
3. Interpretation of coefficients and results, goodness of fit, prediction, etc.
Case:
Poisson flow analysis of customer queuing system, hypothesis testing of product quality, etc.
Analysis of purchase intention of customers of different ages and genders, analysis of the number of failures of food and beverage machines, etc.
The third stage : principal component analysis and time series model
1. Principle, dimensionality reduction and comprehensive evaluation method of principal component analysis
2. Time series data preprocessing, stationary time series inspection method, difference
3. Autocorrelation and partial autocorrelation coefficients, model identification, parameter estimation
4. Model verification, model optimization, model prediction and analysis, etc.
Case:
In order to study the sales of a certain product of a certain retailer, a retailer collected monthly sales data, and based on this data, it made time series analysis and prediction