Smart-Lead-Scoring-Engine

Analysis on finding potential customers to target for lead generation

Problem Statement:

Given problem statement is supervised learning binary classification problem. Here we have 18 independent feature and 1 dependent feature. Training data has 39161 observations and 19 features. We used multiple classification model with best one came out as XGboost Classifier with F1 score as 68%.

Steps Perfomed:

1. Data Preprocessing and EDA- a. Analysing Basic Metrics - shape, datatypes, statistics of the dataset. b. Non Graphical Analysis - value_counts, unique, nuniquue c. Graphical Analysis - HistPlot, KdePlot, Distplot, BoxPlot, BarPlot, CountPlot d. Missing Values and Outliers Detection - BoxPlot and Inter Quantile Range
2. Applied Data Science- a. Check Distribution b. Feature Transformation c. Missing Values Imputation d. Finding Correlation e. CLT and Confidence Interval
3. Modelling- Tries various supervsied learning algorithms like logistic, decision trees, random forest, boosting, bagging, etc. Best Result was given by xgboost with F1 score of 68%.
4. Performed Feature Selection using Feature_Importance attribute of the above tree classifier.