I'm sure we all have multiple aliases and titles that others attach to us throughout our careers. There's a saying that goes something like, your career is in the eye of the beholder. I guess that is the beauty, haha. I don't quite know what mine is yet, but gravitationally there is this inert sense of pull from creating technology that makes a profound difference in someone's life. A tool that helps someone cross a a deep foresty mountainous terrain, and lets them see it clearly, openly, and gain a deeper connection to ones inner self. New ways to think and perceive the tangible, beyond what sits at the edge of vision, reaching what we can't see but still feel.
However, to not get too philosophical, I set out to learn machine learning, deep learning, as far into the iceberg as I can go. What is truly exciting is not knowing where this will lead me. The work might not take me far, but, willingly and hopefully it will help me grasp how systems work, and how we connect to technology and everything around us, all the way down to the bottom of the ocean, to the trough of this so-called iceberg.
Just to prephis all learning and code experiments are done without the aid of a.i to truly understand what is going on under the hood.
Project 1 (Comparing text passages)
First larger project - Whilst learning the fundamentals in python, creating a system that compares texts and creates a statistic report, completed this withput use of numpy and pandas as i wanted to understand what the shortcut syntax used in those modules were actually doing for example getting averages of word lengths or word count in each texts or manually adding tokens to a dictionary instead of pandas doing this automatically.
Project 2 (Data Cleaning)
We then moved on to using pandas to organize and calculate kpi values for a coffee shop from a large data set from kaggle, learning to load csv files using methods that instead of killing program if file can’t be found instead gives user an error message prompting for a check.
Using a config file in which main python file uses so makes change globally, By pulling these out into a separate JSON file, the pipeline code stays generic and reusable. Using unique function to give a list of names and payments i was able to make these maps that allowed me to alter incorrect naming methods in the csv. Pandas has a load of nifty little shortcuts which knocked the code down substantially, using the pd to numeric to convert from string to int so i can perform calculations on the data, is just one of the examples of methods i used, then saving the cleaned csv to a new file so it is readily usable for perfoming calculations on.
Project 3 (Linear regression model)
Our first machine learning model utilised a linear regression model to calculate a prediction for houses price in a number of years all dependant on square footage, age of house, where the house is located e.t.c, all these previous pieces of data on neighbouring houses in different areas allowed me to create a train test split which in this case i think i used an equal train test split, this project is where i probably learnt the most in terms of utilising what i had learn previously like cleaning up data and in this case for example splitting data into seperate columns like city,rural,suburbs all examples of working with and handling data.