Arman Hajisafi

DATA ANALYST at UNSW (University Of New South Wales)
I uncover meaningful insights hidden within vast datasets.
Equipped to enhance DATA-DRIVEN decision-making with confidence.

Read More
My LI Page

About Me

Chapter 1 "Back in time"

After finishing university and getting my CE degree, I started my career as a fresh-faced software developer in the tech industry. Along the way, I explored different programming languages and even tried my hand at being a scrum master. But guess what? I eventually found my true calling: working with data and coding. That's when I discovered Python, and it was love at first code! 😊

Chapter 2 "Today"

As a Senior Data Analyst, I work with the University of New South Wales (UNSW) PVCESE Insights Team, specializing in the ETL process and managing large-scale datasets from diverse sources (internal and external), such as QILT SES datasets. These datasets are the most comprehensive government-endorsed surveys across the student life cycle, spanning from commencement to employment. I work simultaneously in both local and cloud-based environments, which has helped me enhance my programming skills in 'Python and R', alongside gaining practical expertise in Azure cloud services, with a focus on Azure Databricks. Visualization also plays a critical role in my duties, where I utilize PowerBI and Python visualization libraries to transform complex data into clear, actionable visual narratives.

Read More

Say Hello to Bokeh

Bokeh has been around for years but I only recently really discovered it and it didn’t take long to become my favorite Python visualization library.

Plots made with Bokeh are flexible, interactive, and shareable.

Bokeh at a Glance 👇
Flexible
Interactive
Shareable
Productive
Powerful and Open Source

Here is a Quick tips and example of exploring data visulization features using bokeh by Payal Patel " Data Scientist at IBM "

Continue Reading
first service
second service

Don’t Run Loops in Python, Instead, Use These!

Loops come to us naturally, we learn about Loops in almost all programming languages. So, by default, we start implementing loops whenever there is a repetitive operation

However it's important to note that Loops are insanely expensive in Python and if you wish to work with a large number of rows then it's crime

Continue Reading

Using Seaborn

An interesting project just recently done, The correlation between countries population and their world cup performance in 2008 World Cup

Job Ad Dashboard

recently finished this super cool project, a versatile dashboard that enables users to search for open job positions in the market and filter them based on specific skills or tools mentioned in the ads, To develop this project, I utilized a range of technologies including Python, Beautiful Soup, Selenium, Regex, Pandas, Streamlit, Seaborn, Plotly, dockerized and deployed on GCP with an automated CI/CD pipeline on GitHub to ensure seamless updates.

Using Seaborn

Seaborn is great for making less-code pair plot, look at this famouse Iris dataset, which lists measurements of petals and sepals of three iris species

Facet Grid

In the Seaborn library, multiple graphs can be created using the FacetGrid() class. Sometimes the best way to view data is via histograms of subsets. Seaborn's FacetGrid makes this extremely simple.

Using Seaborn

Another cool stuff with joinplot, even automatic regression and kernel density estimation can be performed using the joint plot.

Manipulating the legend

For the given visualization, the legend defaults are not always adequate. And, similar to what we did here, we'd like a legend that describes the scale of the point sizes.

Bubble plots in plotly

A bubble chart is a scatter plot in which the markers are sized based on a numerical variable. You can create bubble plots in Python with the scatter function from plotly express. having some cool features like "Color based on the size of the bubble".

Heatmap

The dataset pertains to an airport's monthly foot traffic for each year, from 1949 to 1960. The figures show how many people (in thousands) used the airport. in this case heatmap is the best choice.

Data Preprocessing

Algorithms and analyses are only as good as the data they're based on. Organizations estimate that nearly 30% of the data they use is inaccurate. Companies lose more than just money when they use this soiled data; it costs them 12% of their overall revenue.

Using Seaborn

Joint plot can be used to visualize the relationship between two variables as well as the distribution of individual values for each variable.

SQL

Data analysis, in many circumstances, involves dealing with huge volumes of data stored in relational databases. Working with such huge volumes of data needs high-level solutions to manage it other than the usual spreadsheets. As the volumes of datasets increase, it becomes untenable to use spreadsheets. The best solution for dealing with huge datasets is SQL. SQL has the capacity to manage such datasets.

Coming Soon