(Random Forest Model to Classify Sales Deals Outcomes)
This project uses IBM data for various sales deal and their characteristics to classify the outcome of sales. This resulted in a Flask app (instructions below).
How can we help sales representatives prioritize deals?
01 Data Acquistion
02 Preprocessing & EDA
03 Modeling
04 Data Visualization
- Pandas
- Numpy
- Matplotlib
- Jupyter Notebook
- Flask
Sales Outcome - Won or Loss
| ID | Name | Description | Type |
|---|---|---|---|
| oppID | Opportunity Number | A unique generated number assigned to the opportunity | Categorical |
| subGroup | Supplies SubGroup | Reporting supplies subgroup. Values are: Batteries & Accessories, Car Electronics, Exterior Accessories, Garage & Car Care, Interior Accessories, Motorcycle Parts, Performance Parts, Replacement Parts, Shelters & RV, Tires & Wheels, Towing & Hitches | Categorical |
| mainGroup | Supplies Group | Reporting supplies group Values are: Car Accessories, Car Electronics, Performance & Non-auto, Tires & Wheels | Categorical |
| region | Region | Name of the Region. Values could be : Mid-Atlantic, Midwest, Northeast, Northwest, Pacific, Southeast, Southwest | Categorical |
| route | Route to Market | The opportunities’ route to market: Fields Sales, Other, Reseller, Telecoverage, Telesales | Categorical |
| daysinstage | Elapsed Days In Sales Stage | The number of days between the change in sales stages. The counter is reset for each new sales stage | Numerical |
| oppOutcome | Opportunity Results | A closed opportunity is won or loss. | Categorical |
| stageChanges | Sales Stage Change Count | Actually a count of number of times an opportunity changes sales stages (back and forwards) | Numerical |
| daysToClose | Total Days Identified Through Closing | Total days the opportunity has spent in Sales Stages from Identified/Validating to Gained Agreement/closing | Numerical |
| daysToQual | Total Days Identified Through Qualified | Total days the opportunity has spent in Siebel Stages from Identified/Validating to Qualified/Gaining Agreement | Numerical |
| oppValue | Opportunity Amount (USD) | Sum of line item revenue estimates by sales representative in American currency | Numerical |
| sizeByRev | Client Size by Revenue Client | size based on annual revenue: 1: < $1M, 2: [$1M, $10M], 3: [$10M, $50M], 4: [$50M, $100M], 5: ≥ $100M | Categorical |
| sizeByEE | Client Size by Employee Count | 1: < 1, 2: [1K, 5K], 3: [5K, 10K], 4: [10K, 30K] , 5: ≥ 30K | Categorical |
| pastSale | Revenue From Client Past Two Years | Revenue identified from this client in past two years: 0: 0, 1, [1, 50K), 2: [50K, 400K), 3: [400K, 1.5M), 4: ≥ 1.5M | Categorical |
| Competitor | Competitor Type | An indicator if a competitor has been identified Values: Known, Unknown, None | Categorical |
| ratioIDtoT | Ratio Days Identified To Total Days | Ratio of total days the opportunity has spent in sales stage: Identifiedover total days in sales process | Numerical |
| ratioValtoT | Ratio Days Validating To Total Days | Ratio of total days the opportunity has spent in sales stage: Validating over total days in sales process | Numerical |
| ratioQualtoT | Ratio Days Qualified To Total Days Ratio | total days the opportunity has been spent in sales stage: Qualified/Gaining Agreement over total days in sales process | Numerical |
| dealsizeCat | Deal Size by Category | Categorical grouping of the opportunity amount (OpportunityAmountUSD): 1: < 10K, 2: [$10K, 25K], 3: [$25K, $50K], 4: [$50K, $100K], 5: [$100K, $250K], 6: [$250K, $500K], 7: ≥ $500K | Categorical |