Excel Minis

Overview

Cleaned and explored a dataset on AI's impact on student learning outcomes — cleaned data, created pivot tables and then produced some analysis. Click download below for the Excel file

Insights

• The Stem field saw the highest weekly use of AI.
• Advanced users of AI (so correct prompting) saw the highest skill retention from using it (82.73%).
• Dependency of AI showed a clear association with increased anxiety from exams.

Overview

This data was collected from investing.com, covering BTC/USD, XAU/USD and SPY/USD over five years from January 2021 to January 2026. A strong caveat: these results are strictly tied to this dataset, the timeline above, and the methodology used. This data should not be used for any financial predictions.

Asset Returns

Over the data timeline (Jan 2021 – Jan 2026):

• Gold shows the highest return over five years by approximately +180% from this data.
• S&P shows the lowest return over five years, as measured by index change.

Indexed Asset Performance

To accurately compare value increases across Bitcoin, gold and the SPY ETF, indexing provides a useful basis.

• This data suggests Bitcoin was by far more volatile, seeing the biggest and longest drop of value between 05/01/2022 to 03/01/2024 whilst seeing a shorter, more drastic increase than gold between the period of 11/01/2024 to 09/01/2025.
• Gold is shown to maintain a steady increase however, running ahead of Bitcoin's activity.
• SPY (S&P500 ETF) and BTC ended at a similar level, with SPY slightly higher.

Asset Volatility

• Bitcoin showed the most dramatic volatility. The steepest dips occurred in May 2021 and January 2022 — Gold and the S&P 500 saw similar downturns but with far less extreme swings.

Overview

This dataset consists of 9800 rows of sales data from a superstore. It is unknown whether the data is synthetic. It shows customer activity on purchases across different categories over a 3-year period.

Insights

• The data suggests that the top product by far in sales is the Canon ImageCLASS 2200 Advanced Copier.
• Within this data, California was responsible for the most sales, while North Dakota had the least. California is one of the largest US states, so this could explain that behaviour.
• Sales across categories are relatively even, with technology leading at 36.6% of sales.
• Sean Miller generated the most sales.
• The category breakdown in each state had a few outliers, but they contributed very few sales respectively, skewing the breakdown.
• Most customers bought from more than one category, at 97.1%.
• Office supplies fluctuate more than other categories, which makes sense given it is the weakest performing category. 01/01/2016, for example, showed a very poor month for office supplies according to this data.