📊 Superstore Dataset — Deep Analysis Notebooks

This repository contains three separate, deep-dive Jupyter notebooks based on the Superstore.csv dataset (9,994 orders — US retail sales data). Each notebook focuses on only one library so you can deeply learn Pandas, NumPy, and Matplotlib individually. All notebooks are fully executed with embedded outputs and charts.

---

📁 Files

Notebook	Focus	Cells
Superstore_Deep_Analysis.ipynb	🐼 Pandas — data cleaning, exploration & business analysis	89
Superstore_NumPy_Analysis.ipynb	🔢 NumPy — array-based numerical computing & stats	83
Superstore_Matplotlib_Analysis.ipynb	📈 Matplotlib — professional data visualization	61

---

🐼 1. Superstore_Deep_Analysis.ipynb (Pandas)

Only Pandas is used to load, clean, explore, and analyze the dataset (with minor Matplotlib support for basic visualization).

Topics covered:

• Data load & inspection — read_csv, info, describe, dtypes, memory_usage
• Cleaning — isnull, duplicated, astype, to_datetime, fillna, rename/drop
• Filtering & indexing — loc/iloc, query, boolean masking
• Sorting — sort_values, nlargest/nsmallest
• GroupBy & aggregation — groupby, agg, transform, filter
• Reshaping — pivot_table, crosstab, melt, stack/unstack
• Transform — apply, map, str/dt accessors
• Binning — cut, qcut
• Stats — corr, rank, cumsum/cummax
• Time series — resample, rolling
• Combine — merge, concat, idxmax/idxmin
• Export — to_csv
• Visualization — pandas .plot() (bar, line, pie)

---

🔢 2. Superstore_NumPy_Analysis.ipynb (NumPy)

The CSV is parsed using Python's built-in csv module (due to quoted fields), but all computations use pure NumPy arrays — zero Pandas involved.

Topics covered:

• Array basics — np.array, dtype, shape/ndim/size, reshape/ravel/transpose
• Indexing — slicing, boolean masking, fancy indexing, np.where
• Sorting — np.unique, np.sort/argsort
• Statistics — mean/median/std/var, argmin/argmax, percentile, ptp
• Group-aggregation without pandas — boolean masks + np.bincount
• Cumulative — cumsum, cumprod, diff
• Correlation — np.corrcoef, np.cov
• Broadcasting — profit margin %, discount simulation, clip/round/abs/log/sqrt
• Distribution — np.histogram, np.bincount
• Regression — np.polyfit, np.linalg.lstsq (trend line)
• Combining arrays — concatenate/vstack/hstack/split
• Random — np.random bootstrap resampling (95% confidence interval)
• NaN handling — isnan, nanmean/nanstd/nansum
• Date/time arrays — datetime64, timedelta64
• Structured (record) arrays
• Save/Load — np.save, np.savetxt, np.load

---

📈 3. Superstore_Matplotlib_Analysis.ipynb (Matplotlib)

Data preparation is done using NumPy and the csv module, but every chart and customization is built purely with Matplotlib.

Topics covered:

• Figure/Axes anatomy — plt.figure, plt.subplots
• Line plots — multi-line, styles, markers
• Bar charts — vertical, horizontal, grouped, stacked
• Histograms — normal, cumulative, overlapping
• Scatter plots — basic + bubble chart (color/size mapping)
• Pie charts
• Box & violin plots
• Layouts — subplot2grid, GridSpec
• Twin axes — twinx()
• Annotations — annotate, text, arrows
• Colormaps & colorbar
• Heatmaps — imshow
• Contour plots — contour/contourf
• Stem & step plots
• Error bars & fill_between
• Date axis formatting — matplotlib.dates
• Log scale axes
• 3D plotting — scatter + bar3d
• Styles & rcParams customization
• savefig (high-res export)
• Final combined executive dashboard

---

📌 Dataset Info

Superstore.csv — Order-level retail dataset from a US superstore (2014–2017). It contains 9,994 rows and 21 columns, including Order/Ship dates, Customer details, Product Category/Sub-Category, Sales, Quantity, Discount, and Profit.

👤 Author

Azeem — @Azeem8541