saintuche/pharmacy-sort

Uses data from Oriel to Rank pharmacies acrosss the UK based off pay and location filters. Used by friends at University of Nottingham to select their placements and was very succesful

Pharmacy-sort

Uses data from Oriel to Rank pharmacies acrosss the UK based off pay and location filters. Used by friends at University of Nottingham to select their placements and was very succesful!

Motivation

A friend informed me that she had only 6 days left to decide her post-grad pharmacy placement through the website oriel. Unfortunately, the website did not have any filters and you would have to scroll down a list of 3000 pharmacies. Furthermre, information was only shown on the pharmacy when it was clicked on so each indivdual pharmacy had to be clicked. Thhey could choose any number of pharmacies but most people seemed to choose around 50. So as you can imagine, this was a very long task, hence why they were given 3 months to do it. However, as most university students do, she procrastinated left it to the last minute so my aim was to help her get it done much faster.

Goal

My goal was to generate the list for her automatically based off her preferences. The two things that concerned her were pay and distance.

The Data

There was an excel download of the list of pharmacies (Preferences.xlsx). This contained all the data of the 3000 pharmacies stored in a table. I needed to get the postcode and the Salary of each pharmacy. The postcode is stored within the pharmacy title. So postcode had to be extracted from the name. Using get_postcode in choice.py. The salary was stored in salary column so all I had to to was read through each row.

Approach

First I had to decide which mattered more and how much more between pay or distance. I decided to use a weighted scoreing system which the user decides. a salary weight of 60% would mean distance has a weight of 40% which means salary matters more to the user by 20%. Once that is decided A way to calculate them would be needed. Calculators.py does exactly this.

Calculators.py

salary_score(salary, max, min) -> returns a score out of 100 based off the users minumun accepted salary and the maximum available. If the salary is less than the minimum 0 is returned, if the salary is exactly the minimum 1 is returned and if is exactly the maximum 100 is returned.

distance_score(goal_location, start_location, max_distance) -> returns a score based of max distance. Uses google api to take in two post codes and calculates distance.

distance_score_to_distance(score, max_distance) -> if distance score is greater than 0 the distance of the pharmacy is returned. This means if the distance is within the users maximum travel distance a score will be generated. A score of 100 would mean the user does not have to travel at all and a score of 1 would mean the user would have to travel exactly their maximum prefered distance.

def total_score (salary_score, distance_score, salary_weight, distance_weight) -> calculates weighted mean of the job taking into accoung distance and salary

Map.py

Uses google maps API to calculated distance between two postcodes in miles. Needs google API key.

Choice.py

main file that generates list as new sheet in excel

Outcome

The program was succeful and was very popular amongst her peers. It cut the time of selecting pharmacies down from months to just a few minutes. Head of deparment was also interested in the code.

Future improvements

May work on improving later Pandas could have been used to read data files instead. Concurrency programming could have been used as it take around 10 minutes to generate list

Made in November in 2020, uploaded in 2022