In this note, I will tell you how to deploy a model for free up to a certain level of use and not bother with writing a microservice. I note that such a solution is easily integrated, for example, into a web service. All you need is to use Google Cloud Functions.
Google Cloud Functions is a serverless approach, i.e. server services are provided without renting or purchasing equipment. With this approach, the provider manages infrastructure resources, configures and maintains them.
The main advantage of Google Cloud Functions is automatic scalability, high availability and fault tolerance.
What is Monthly Recurring Revenue?
Monthly Recurring Revenue – regular monthly income. This metric is used primarily in subscription models. In this case, the income itself must be reduced to months.
Why is it valuable?
If we have a subscription service, we have regular or periodic payments, then we can understand how much money we will earn and how effective our business is. Further, we can increase MRR by switching customers to a more expensive tariff or try to reduce customer churn.
For this task, use the new dataset: https://alimbekov.com/wp-content/uploads/2021/03/mrr.csv
customer_id – already familiar customer ID
first_order – Subscription start date
EndDate – Subscription end date
rate – subscription plan (monthly, semi-annual, annual)
Amount – amount paid
commission – payment system commission
We will use the following formula to calculate MRR: MRR = new + old + expansion + reactivation – churn – contraction
- new MRR – the first payment of a new client
- old MRR – recurring customer payment
- expansion MRR – increase in MRR due to the new tariff
- contraction MRR – decrease in MRR due to the new tariff
- churn MRR — MRR outflow due to termination of payment
- reactivation MRR – return of a client who had an outflow of MRR
What is cohort analysis?
Cohort analysis consists in studying the characteristics of cohorts / vintages / generations, united by common temporal characteristics..
A cohort/vintage/generation is a group formed in a specific way based on time: for example, the month of registration, the month of the first transaction, or the first visit to the site. Cohorts are very similar to segments, with the difference that a cohort includes groups of a certain period of time, while a segment can be based on any other characteristics.
Why is it valuable?
This kind of analysis can be helpful when it comes to understanding the health of your business and the stickiness of your customers. Stickiness is critical, as it is much cheaper and easier to retain a customer than it is to acquire new ones. Also, your product evolves over time. New features are added and removed, design changes, etc. Observing individual groups over time is the starting point for understanding how these changes affect user/group behavior.
Make an RFM analysis. It divides users into segments depending on the prescription (Recency), frequency (Frequency) and the total amount of payments (Monetary).
- Recency – the difference between the current date and the date of the last payment
- Frequency — number of transactions
- Monetary – amount of purchases
These three indicators must be calculated separately for each customer. Then put marks from 1-3 or 1-5. The wider the range, the narrower segments we get.
Points can be set using quantiles. We sort the data according to one of the criteria and divide it into equal groups.
For this task, we use the public dataset: https://www.kaggle.com/olistbr/brazilian-ecommerce nd the olist_orders_dataset.csv and olist_order_payments_dataset.csv files. You can connect them
The field of medical imaging has become very popular in recent years. Therefore, I write book where you will learn the basics of medical image analysis using Python. You will study CT and X-ray scans, segment images, and analyze metadata. Even if you have not used with medical imaging before, you will have all the necessary skills upon completion of the book.