Access to data

Top 20 data tips for ML

1. Data Is paramount in ML . What feature vectors can we rely on to build your model ?
2. We are typically starved on data; e.g. product , customer, customer preferences, transactional Data.
3. Context aggregation between various shopping patterns opens up numerous possibilities .
4. Segment clients based on aggregation of context – micro segmentation.
5. Retailers are generally wary of transactions  falling in hands of competitors to steal customers . 
6. Retailers and financial institutions are sensitive to consumer and customer privacy . The anti creep norm .
7. Create micro segments for highly engaged segments .
8. Txn based or product based vs segment based is the preferable way of leveraging contextual relevance .
9. Use Social data to enhance segments  through context aggregation .
10. Retailers are overwhelmed  on how to act on data
11. Acq of customers through increased insight . Engaged clients of an industry increase revenue . Capture their attention .
12. Enrich to understand what is here Ing on with input of social handles 
13. Life events detection
14. Models . Enrich visual models 
15. Extraction emotional memes and how they change over time 
16. Correlate between structured data elements 
17. Combine structured and not n structured data to form compelling business insights 
18. Detect events such as Party or vacation on instagram etc 
19. Opt in for Sharing social media 
20. Start journey of data exploration with Existing data 
21. Get clients to Opt in to provide social media handles and transactions 
22. Aggregation is /brings increasing insight when the pieces of the puzzle are brought together in context .
23. Predictive models are high along the maturity spectrum right below ml and deep learning models