programmer
Uber Movement SANRAL
Keywords
Skills
R, Data science, Machine Learning, Forcasting, Statistics, Data-Mining, SQL developer, SAS, Guide/Miner, Clustering, data cleaning, Data Analysis, K-means, SQL, Machine Learning, Supervised learning, Regression, SVM, Decision Tree, Unsupervised Learning, Data Science, Oracle DB, Toad, Analytics
Project history
01/2020
-
09/2020
* Ranked TOP 11% '13th ' position (only 113 succeed to submit among 711 competitors)
* The objective of this competition is to predict when and where road incidents will
occur next in Cape Town
* The objective of this competition is to predict when and where road incidents will
occur next in Cape Town
01/2018
-
03/2020
Data Scientist
Groupe Tunisie Telecom
CVM and Data-Mining department (SQL developer & SAS (Guide/Miner) user) :
* Clients Segmentation (Clustering & Classification) ::
* Data preparation: creating of the ABT, ensuring data quality, data cleaning
(imputation and transformation ...)
* Data Analysis : detailed exploratory analysis including correlation and different tests
* Clustering : identifying clusters of customers who have the same behavior within the
same group of clients (using K-means), detailed profiling of the segments for
validation
* Classification : assigning new customers to the clusters that have been created (using
decision Tree)
* The model was built with SAS (Guide and Miner)
* Multi_Sim Users Detection ::
* The goal is be able to identify from some pre-hypothesis, the users who have two sim
providers, include testing the hypothesis, detailed profiling of the target, giving a final
score to each client.
* Model was built with SAS (Guide and Miner)
* Family Communities ::
* From a predefined communities results, being able to distinguish into each
community, those who have a family relation (the big family and those who lives int
the same house) and those have not.
* Contributing of building and assisting in a series of workshops along with experts from SAS,
KPMG and Business & Decision, for different models including :
* Churn, Cross-sell and Community Link Analysis (SAS)
* Clients Segmentation (Clustering & Classification) ::
* Data preparation: creating of the ABT, ensuring data quality, data cleaning
(imputation and transformation ...)
* Data Analysis : detailed exploratory analysis including correlation and different tests
* Clustering : identifying clusters of customers who have the same behavior within the
same group of clients (using K-means), detailed profiling of the segments for
validation
* Classification : assigning new customers to the clusters that have been created (using
decision Tree)
* The model was built with SAS (Guide and Miner)
* Multi_Sim Users Detection ::
* The goal is be able to identify from some pre-hypothesis, the users who have two sim
providers, include testing the hypothesis, detailed profiling of the target, giving a final
score to each client.
* Model was built with SAS (Guide and Miner)
* Family Communities ::
* From a predefined communities results, being able to distinguish into each
community, those who have a family relation (the big family and those who lives int
the same house) and those have not.
* Contributing of building and assisting in a series of workshops along with experts from SAS,
KPMG and Business & Decision, for different models including :
* Churn, Cross-sell and Community Link Analysis (SAS)
08/2019
-
09/2019
Data Scientist
Text Classification, NLP (R&D)
* The goal is to be built a model that can be used for a future data to be classified into
categories.
categories.
02/2019
-
04/2019
Vehicle Weight Estimation
* From very sensitive sensors values (with milliseconds), being able to predict the
correct weight or in worse case, being able to reduce the maximum error in
difference between real value and the predicted one.
correct weight or in worse case, being able to reduce the maximum error in
difference between real value and the predicted one.
12/2018
-
01/2019
Data Scientist
* Using the page's API, collecting every information about the page; fans, number of
comments, daily visits for certain promoted videos.
* Applying sentimental analysis on the collected comments to see feedback of fans.
comments, daily visits for certain promoted videos.
* Applying sentimental analysis on the collected comments to see feedback of fans.
06/2016
-
12/2017
Data Analyst
Groupe Tunisie Telecom
CVM and Data-Mining department (SQL developer & SAS (Guide/Miner) user) :
* Reports Sheets and Dashboards for various analysis (Monitoring the CVM results, Analyzing
the KPI of the Network quality, Customer detailed profiling and, Sim-card sells, Market
growth, Data penetration ..)
* Automation of the reporting process for PowerPoint presentations via VBA
* Verification of anomalies in the databases
* Recommendation of new offers and retention actions (Targeting campaigns, Try&Buy
offers)
* Reports Sheets and Dashboards for various analysis (Monitoring the CVM results, Analyzing
the KPI of the Network quality, Customer detailed profiling and, Sim-card sells, Market
growth, Data penetration ..)
* Automation of the reporting process for PowerPoint presentations via VBA
* Verification of anomalies in the databases
* Recommendation of new offers and retention actions (Targeting campaigns, Try&Buy
offers)
06/2015
-
12/2015
Research Statistician
LAAS-CNRS
an automatic management of a home security system property (for the Smartfox Project with MyFox) :
* The idea is to develop an automatic model to know from the owner's habits (always the
case of badge absence), if it's the owner who tries to enter or it's an intrusion based on
the owner behavior learned by the model over an historical period, and then decide to
launch the alarm or to deactivate it automatically, it includes ::
* Data analysis
* Statistical learning
* Database management (feeding the server in real time)
* My book (in french) was published by the EUE (Editions Universitaires Européennes)
within this link https://www.morebooks.de/store/es/book/gestion-automatique-
d%E2%80%99un-syst%C3%A8me-de-s%C3%A9curisation-des-biens-%C3%A0-
domicile/isbn/978-3-639-52907-4
* The idea is to develop an automatic model to know from the owner's habits (always the
case of badge absence), if it's the owner who tries to enter or it's an intrusion based on
the owner behavior learned by the model over an historical period, and then decide to
launch the alarm or to deactivate it automatically, it includes ::
* Data analysis
* Statistical learning
* Database management (feeding the server in real time)
* My book (in french) was published by the EUE (Editions Universitaires Européennes)
within this link https://www.morebooks.de/store/es/book/gestion-automatique-
d%E2%80%99un-syst%C3%A8me-de-s%C3%A9curisation-des-biens-%C3%A0-
domicile/isbn/978-3-639-52907-4
Local Availability
Only available for remote work
Yes I'm ready to travel for a specific period of time
