### // Assighment 1

### The Engineering Problem

*(The below are guidelines only, but highly recommended to make sure you are covering the most important points for your manageers.)*

You work for a gas plant and are responsible for

**Atomic mass**

C (carbon) 12 H (Hydrogen) 1 S (Sulfur) 32

N (Nitrogen) 14 O (Oxygen) 16

Molecular weights

DGA: 105 lb/lbmol MDEA: 119 lb/mol H2S: lb/lbmol CO2: lb/lbmol

Densities

DGA: 8.83 lb/USG MDEA: 8.75 lb/USG H2O: 8.33 lb/USG

Concentrations

- 1lb = 7,000 grains
- 1kg = 15,425 grains
- 1wt% = 584.2 grains/US gallon
- 1 ppmv = 0.05835 grains/US gallon
- ppm x 0.001 = grams/liter
- 1 wt% = 700 grains/ imperial gallon
- grain/gallon x 17.12 = ppm
- grain/gallon x 0.01712 = gram/liter
- 1 grain H2S/100 SCF = 16 ppmv H2S
- ¼ H2S/100 SCF = 4 ppmv H2S
- 4 ppmv = 6 mg H2S/ Nm^3

Due to a financial scandal that hit a competitive bank last week, you suddenly have an influx of new people applying for loans for your bank instead of the other bank in your city. All of a sudden you have nearly 500 loan applications to process this week!

Your manager sees this new influx as a great opportunity and wants you to figure out how to process all of these loan applications within one week.

Fortunately for you, you just completed a course in classification modeling and know how to systematically evaluate the creditworthiness of these new loan applicants.

For this project, you will analyze the business problem using the Problem Solving Framework and provide a list of creditworthy customers to your manager in the next two days.

You have the following information to work with:

- What quantity of acid gas (lbmol/hr or kmol/hr) is removed by the solvent if we try to achieve a CO2 specification of 2% and an H2S specification of < 4 ppm?
- Determine the rich loading given the following two amine circulation rates:
- 1300 gpm of 33 wt% DEA
- 1000 gpm of 45 wt% MDEA

- What would on appropriate steam rate be for this required duty?

### Steps to Success

#### Step 1: Engineering and Data Understanding

Your project should include a description of the key engineering decisions that need to be made.

#### Step 2: Explore and Cleanup the Data

To properly build the model, and select predictor variables, you need to explore and cleanup your data.

Here are some guidelines to help you clean up the data:

- Are any of your numerical data fields highly-correlated with each other? The correlation should be at least .70 to be considered “high”.
- Are there any missing data for each of the data fields? Fields with a lot of missing data should be removed
- Are there only a few values in a subset of your data field? Does the data field look very uniform (there is only one value for the entire field?). This is called “low variability” and you should remove fields that have low variability. Refer to the “Tips” section to find examples of data fields with low-variability.
- Your clean data set should have 13 columns where the Average of
`Age Years`

should be 36 (rounded up)

**Note:** If you decide to impute any data field, for the sake of consistency in the data cleanup process, impute the data using the median of the entire data field.

#### Step 3. Train your Classification Models

You should choose 70% to create the Estimation set and 30% to create the Validation set. Set the Random Seed to 1 if you’re using Alteryx.

Train your dataset using these models:

- Logistic Regression
- Decision Tree
- Forest Model
- Boosted Tree

#### Step 4. Writeup

Compare all of the models’ performance against each other. Decide on the best model and score your new customers.

**Important**: Your manager only cares about how accurate you can identify these calculations who and do not qualify for loans for this problem.

Write a brief report on how you came up with your classification model and write down how many of the new customers would qualify for a loan.Assignment