Objectives of the analysis:
1. To identify the factors that can predict various Facebook metrics (especially different kinds of user engagement)
2. To develop predictive models for the Facebook metrics
3.
Copyright By PowCoder代写 加微信 powcoder
Minimum requirements:
1. Linear regression model
2. 2 machine learning models (e.g. neural network model, random forest, SVM)
3. Principal components analysis of the Facebook metrics
4. Cross-validation with 70% training and 30% validation set
5. Suggest another set of Facebook metrics that can be included in the analysis to gain more insights
Suggested data preparation steps:
· Add attributes for Company ID, before and after introduction of emojis, before and after introduction of live streaming
· Don’t process the data using MS Excel, as MS Excel will mess up the dates and numbers.
· Read the .csv file directly into R Dataset. Use read.csv
· To split the status_id into company_ID and record_ID, you can use the string function str_split_fixed. Consult the stringr cheatsheet (attached).
For example: Dataset$company_ID <- str_split_fixed(Dataset[,1], "_", 2)[,1]
· To extract the year from status_published, you can also use str_split_fixed.
For example: str_split_fixed(Dataset$status_published, "[ /:]", 5)[,3]
· If you want to store status_published as a Date (in order to do date operations), use as.Date.
For example: as.Date(Dataset$status_published, tryFormats = c("%m/%d/%Y"))
程序代写 CS代考 加微信: powcoder QQ: 1823890830 Email: powcoder@163.com