I have been self-learning data science from different sources. I have a dataset which was sent to me by my friend from one of her college courses. The work is to be done in R preferably.
Description: It is the response of survey and results regarding usage of services for financial transactions using traditional methods or mobiles as a medium in a developing country. The study was conducted to understand the usage of technology for financial services by men and women in developing countries to find out how comfortable they are to use technology for this purpose. A dictionary is provided with the dataset which you may refer to column details. The dataset is big and contains many missing values. In the dataset, for male they have used 0 and for female are using 1. The purpose of this problem is that you will be able to understand how to process real data for valuable information extraction.
Dataset dictionary: https://1drv.ms/x/s!ArGTGzC7esWNa25KnJACZTXuGPw
The tasks given to my friend were:
1) Preprocess the data
2) Which feature engineering techniques you used?
3) predict the gender of the survey respondent.
4) Predict that respondent is female with a certain probability. 5) Find interesting patterns that are associated with a specific gender
1) I want to know what preprocess can I do in the data. I can remove missing and duplicate values, but for missing values there is a problem. If the survey respondent chose other in the religion column (96 value in column DG3A), they specified the name in the next column. But if they gave a response such as 1 (which is for Christianity) in the DG3A column then the value in the next column would be empty. How would I cater for these missing values (which aren't really missing instead needed not to be filled)?
2) I am not sure which ones should I use and how.
Which anyone can also guide me which columns to use and how for the remaining parts it would be great. :)