Medication is often expensive, across all areas of treatments. It is often difficult to tell if something will work for you and be worth the price. Everybody is unique, and that poses a challenge in that it can be difficult to tell what will work for another and not you. Of course, your healthcare provider knows you best, but with so many options regarding a singular medication's form, price, and ease of use, what is the best option?
This dataset contains common treatments for 37 different conditions. The goal here is to determine a medication's satisfaction rate based on price, form, ease of use, and reviews. With this, we can find a medication type that has worked the best for the most people.
If succesful, this project will yield a prediction of patient satisfaction with a medication. This predictor will provide a useful tool in allowing both patients and manufacturers to take satisfaction rate into consideration before either distributing or purchasing medication.
import pandas as pd
import numpy as np
data = pd.read_csv('Drug_clean.csv')
data.head()
Condition | Drug | EaseOfUse | Effective | Form | Indication | Price | Reviews | Satisfaction | Type | |
---|---|---|---|---|---|---|---|---|---|---|
0 | Acute Bacterial Sinusitis | Amoxicillin | 3.852353 | 3.655882 | Capsule | On Label | 12.590000 | 86.294118 | 3.197647 | RX |
1 | Acute Bacterial Sinusitis | Amoxicillin-Pot Clavulanate | 3.470000 | 3.290000 | Liquid (Drink) | Off Label | 287.370000 | 43.000000 | 2.590000 | RX |
2 | Acute Bacterial Sinusitis | Amoxicillin-Pot Clavulanate | 3.121429 | 2.962857 | Tablet | On Label | 70.608571 | 267.285714 | 2.248571 | RX |
3 | Acute Bacterial Sinusitis | Ampicillin | 2.000000 | 3.000000 | Capsule | On Label | 12.590000 | 1.000000 | 1.000000 | RX |
4 | Acute Bacterial Sinusitis | Ampicillin | 3.250000 | 3.000000 | Tablet | On Label | 125.240000 | 15.000000 | 3.000000 | RX |
We will use a Kaggle data set of Drug Performance Evaluations to observe the features of each drug:
We will cluster the medications into groups based on satisfaction rates, and from there we will have a grouping of what attributes these medications have. This will seperate medications based on how well they support patients, rather than what they specifically treat in order to gain insight on effectivness.