Can you foresee a customer's next move?
The Boston Institute of Analytics (BIA) proudly presents a collection of student presentations on data analysis projects exploring customer transaction prediction.
Embark on a journey to uncover the secrets of anticipating customer behavior. These presentations showcase innovative approaches to analyzing customer data and predicting future transactions, providing valuable insights for:
Retail and e-commerce businesses looking to optimize inventory management and targeted promotions
Marketing professionals aiming to personalize customer experiences and boost conversions
Data analysts and enthusiasts seeking to learn cutting-edge customer behavior analysis techniques
This compelling collection offers:
visit https://bostoninstituteofanalytics.org/data-science-and-artificial-intelligence/
In-depth exploration of data analysis methods for transaction prediction
Real-world case studies demonstrating the power of predictive analytics
Insights and findings from the research of talented BIA students
A springboard for developing your own customer transaction prediction strategies
Gain a competitive edge by harnessing the power of data. Watch these presentations and unlock the secrets of customer behavior prediction today!
visit https://bostoninstituteofanalytics.org/data-science-and-artificial-intelligence/
3. By Atharva Kulkarni
What is Santander ?
Santander is an American bank operating as a wholly-
owned subsidiary of the Spanish Santander Group. It is
based in Boston and
its principal market is the Northeastern United States.
Why this model is
important?
To identify who will make a
transaction.
What will be the Impact of
Model?
The model will predict and help Santander with the problem of
identification of the customers who will make a transaction with
the bank in future.
4. By Atharva Kulkarni
WORK FLOW
Data
collection
Exploratory
Data Analysis
(EDA)
Preprocessing
Visualization
Dividing Data
into X and Y
Model
selection and
Evaluation
First importing necessary libraries like Pandas, Seaborn,
Matplotlib.pyplot and Numpy.
5. By Atharva Kulkarni
DATASET
Santander Customer
Transaction Prediction
This Dataset Consist of Two Files
Train Data Test Data
Rows Columns Rows Columns
200000 202 200000 201
Null Values
Train and Test Data does not
contain Null Values
Duplicate Values
Train and Test Data does not
contain Duplicate Values
6. By Atharva Kulkarni
Exploratory
Data Analysis
(EDA)
Exploratory Data Analysis (EDA) refersto the method of studying and exploring record sets to apprehend
their predominant traits, discover patterns, locate outliers,and identifyrelationships betweenvariables.
EDA is normally carried out as a preliminarystep beforeundertakingextra formalstatistical analyses or
modeling.
df.head() = Display first Five Rows of Dataset
df.tail() = Display last Five Rows of Dataset
df.describe() = Gives descripive Statistics of Datadet
df.isnull().sum() = Display the number of Null Values in Dataset
df.shape() = Display the number of Rows and Columns of Dataset
df.dtypes() = Display the Data Type of each Feature of Dataset
df.info() = Gives the Summary of Dataset including column name, data type,
non-null values and memory usage
11. By Atharva Kulkarni
Dividing Dataset into
‘X’ and ‘Y’
Split the Data into Training Data and
Testing Data
The Data is imbalanced so
we used UnderSampling to
Balanced the Data
12. By Atharva Kulkarni
Model1 = RandomForestClassifier
Accuracy Score
Model2 = LogisticRegression
Accuracy Score
MODEL Selection
15. By Atharva Kulkarni
CONCLUSION
● Our model can be used to find the right customers to target and
increase profits, as well as return on marketing investment.
● After dealing with data imbalance our data was ready for feature
engineering.
● Best Method – The Algorithm KNeighborsClassifier Suits best for my
dataset because it gives accuracy of 0.891825.