Cashflow Analysis.
Ask the simple questions.
We wanted to know what the future of the organization would look like. Instead of using Excel like sane people, we decided to try building a boutique solution from the ground up. Well, you can scroll down to see how it went.
PS: we had some fun with the white labeling, enjoy.
Working Process
How we build a Cashflow Model.
We decided the best way to showcase how we build a cashflow model is to do it for ourselves and see what we learn. Read along to see what we did. If you want us to build something like this for you get in touch.
Let's Connect1-6 May, 2024
Problem Definition
Defining the problem is usually half the battle won, so we started off with pulling together all the stakeholders with the aim of defining the problem statement and data sources we could use.
6-9 May, 2024
Data Exploration
Your model is only as good as your data, So before we got to building we took a look at all the possible sources of data and made sure we had enough data so we could make proper estimations and predictions.
10-17 May, 2024
Hypothesis validation
Once we concluded we had enough data we started with build a simple MVP version and back tested the results to ensure that it predicted the results with an acceptable level of accuracy and fitting the general trend.
17-24 May, 2024
Expectation setting with stakeholders
Before we started the build, we informed all the relevant parties about the results and set expectations about the system. We also asked what all controls they would need, how long before we needed to retrain the model and what all feature were important. this helped us estimate the time and compute required.
24 May-15 June, 2024
System Design
Now That the ground work was laid all we had to do was build the final system. We Designed a blueprint of the architecture and started building it with a ground up approach of building Bae modules which lead to building up of the complete system.
15-28 June, 2024
Fine Tuning
We took the final Model and showed it to the department head, and we reviewed the results and iteratively updated the system fixing some issues with the primary data source[clerical errors haunt every dataset]. and final we have the complete model.
1 July, 2024
Deployment
Finally we took the model after it was properly vetted and deployed it, we white labeled the customer data and deployed the model. Click round and see how our clients performed.
Welcome to the Playground.
If you want to look around you will need to open the site on a desktop! 🎉