You are given a set of Twitter accounts of 10 million Twitter accounts.

1) You are asked to sample 10% of the accounts using different sampling methods. How do you apply “Stratified sampling” and “Sampling without replacement” on the given dataset?

2) (Missing data) You need to work on location data, but many of the users do not reveal their location in their profiles. What approach do you use to deal with missing location attributes in the Twitter dataset?

3) (Duplicate data) Propose an approach to find highly similar Twitter accounts.