agamemnon

Popular articles by agamemnon

Price and promotion optimization for FCMG

Posted 5 CommentsPosted in Learn, Prediction systems, Team solutions

   Introduction Data provided consists of 3 years of weekly volume of sales, price of product in question, prices of main competitors and promotion calendar for a FCMG product. Data is provided by SAP. The task is to identify the volume uplift drivers, measure the promotional effectiveness and measure the cannibalization effect from main competitors. […]

The SAP Case using KNIME and Multiple Linear Regression Method

Posted 5 CommentsPosted in Team solutions

SAP Case Team Mentors:  Agamemnon Baltagiannis SAP Case Team: Abderrahim Khalifa                                                                | Morocco Andrei Deusteanu (andrei.deusteanu@gmail.com) | Romania Julian Borisov (julian.borisov@yahoo.com)        […]

CASE SAP, TEAM 31415

Posted 3 CommentsPosted in Team solutions

About¶ Entry: Data Science society Datathlon 2018 Case: SAP Case Dataset: available here Authors: Hristo Piyankov (hpiyankov@gmail.com) Notes: not all caclulations and graphs are carried out in python, due to time constraints Business understanding¶ Goal of the study is to udnerstand drivers behind sales up-lift with relation to the company’s own pricing strategies, promostions and […]

Tiny smart data modelled with a not-so-tiny smart model – the Case of SAP

Posted 1 CommentPosted in Team solutions

Tiny smart data modelled with a not-so-tiny smart model Introduction Metadata Business Understanding Data Understanding Data Preparation Modelling Evaluation Deployment Conclusion Metadata Case: The SAP Case – Analyze Sales Team: Chameleon Project URL: https://github.com/Bugzey/Chameleon-SAP Memebers: Stefan Panev (stephen.panev@gmail.com), Metodi Nikolov (metodi.nikolov@gmail.com), Ivan Vrategov (ivanvrategov@gmail.com, Radoslav Dimitrov (rdimitrov@indeavr.com) Mentors: Alexander Efremov(aefremov@gmail.com) Agamemnon Baltagiannis (agamemnon.baltagiannis@sap.com) Team Toolset: […]

Antelope SAP

Posted 2 CommentsPosted in Team solutions

The current paper examines the factors that influence the increase of the
sales volume of a retailer. The aim of the study is to create an accurate model with high explanatory
power which accounts for the promotional and competitor effects on the quantity sold as well
as to identify the main volume uplift drivers. That information could be useful when designing
marketing strategies in order to gain a competitive advantage over the other market players.

Popular comments by agamemnon

Price and promotion optimization for FCMG

Excellent work team!
The most difficult part of this challenge is to understand the data, create new features and rerun the predictive models till you achieve a good accuracy.
As you may mentioned if you run a predictive model with the initial dataset you will get an extremely low modelling accuracy.

I will vote based on the below criteria:
1. business understanding
2. feature engineering
3. modelling accuracy
4. insights & final results

You achieved a good modeling accuracy and you created a number of new features based on your excellent understanding of the data and the business case.
Moreover you visualized all the variable in a meaningful way and you took the right decisions on creating new features.

You could further increase the accuracy of the model by implementing a better base price algorithm and by calculating the baseline volume in a better way.

The SAP Case using KNIME and Multiple Linear Regression Method

Great great work team! The most difficult part of this challenge is to understand the data, create new features and rerun the predictive models till you achieve a good accuracy.
As you may mentioned if you run a predictive model with the initial dataset you will get an extremely low modelling accuracy.

I will vote based on the below criteria:
1. business understanding
2. feature engineering
3. modelling accuracy
4. insights & final results

Your approach is on the right path and you managed to create the new features that will help you identify the volume uplift drivers and increase the accuracy of the model. This is a great accuracy for this specific dataset! This comes as a result of your great business understanding of the case and the dataset as well

Tiny smart data modelled with a not-so-tiny smart model – the Case of SAP

Excellent work team!

The most difficult part of this challenge is to understand the data, create new features and rerun the predictive models till you achieve a good accuracy.
As you may mentioned if you run a predictive model with the initial dataset you will get an extremely low modelling accuracy.

I will vote based on the below criteria:
1. business understanding
2. feature engineering
3. modelling accuracy
4. insights & final results

You achieved a fair modeling accuracy and you created a number of new features based on your good understanding of the data and the business case.
You could further increase the accuracy of the model by implementing a better base price algorithm and by identifying the base volume

CASE SAP, TEAM 31415

Great work team!
The most difficult part of this challenge is to understand the data, create new features and rerun the predictive models till you achieve a good accuracy.
As you may mentioned if you run a predictive model with the initial dataset you will get an extremely low modelling accuracy.

I will vote based on the below criteria:
1. business understanding
2. feature engineering
3. modelling accuracy
4. insights & final results

You achieved a fair but not great modeling accuracy and you created a number of new features based on your good understanding of the data and the business case.
You could increase the accuracy of the model by implementing a base price algorithm and then by taking the % of difference between the actual and the base price you could extract the weekly promotional price reduction and use it as input parameters for your regression model. This was your pain point on a very nice job

Antelope SAP

The most difficult part of this challenge is to understand the data, create new features and rerun the predictive models till you achieve a good accuracy.
As you may mentioned if you run a predictive model with the initial dataset you will get an extremely low modelling accuracy.

I will vote based on the below criteria:

1. business understanding
2. feature engineering
3. modelling accuracy
4. insights & final results

Your approach was on the right path but you didn’t managed to create the new features that will help you identify the volume uplift drivers and increase the accuracy of the model.
You understand the data and the business challenge but you didn’t manage to increase the accuracy of the 1st model by using feature engineering.
You could increase the accuracy of the model by implementing a base price algorithm and then by taking the % of difference between the actual and the base price you could extract the weekly promotional price reduction and use it as input parameters for your regression model.