EDA in Public (Part 2): Product Deep Dive & Time-Series Analysis in Pandas

In Half 1, we took a messy, half-million-row gross sales file and obtained it into form. Particularly, we:

• Sampled the info for quicker processing.
• Fastened lacking values and standardised textual content entries (like ‘EIRE’ to ‘Eire’).
• Filtered out all of the noise — returns, cancellations, and transactions with zero value.
• Engineered our most crucial characteristic: the Income column.
• Achieved our first enterprise perception: the Prime 10 Income-Producing Nations.

We now have a clear, revenue-ready Pandas DataFrame able to be exploited for insights! By the top of this collection, I wish to have mastered exploratory knowledge evaluation utilizing Pandas. Be at liberty to observe alongside for those who’re an information fanatic.

Now we shift our focus to delivering really impactful enterprise intelligence for NovaShop. Our total purpose for Half 2 is to reply the elemental questions: What merchandise are the perfect performers, and when is the perfect time to promote them?

This half is all about evaluation utilizing highly effective Pandas aggregation methods (.groupby()) and have engineering utilizing the datetime column (.dt accessor). We’ll break down our evaluation into two key areas:

• Product Deep Dive: Figuring out the merchandise that transfer probably the most items versus those who usher in probably the most money.
• Gross sales Timing (Time-Sequence): Uncovering the seasonality of gross sales (month-to-month peaks) and the operational traits (busiest hour of the day).

With out additional ado, let’s get to it.

Loading the clear knowledge

To enhance efficiency, let’s export our cleaned dataset and import it once more for evaluation

df.to_csv('online_retail_clean.csv', index=False)

Now, I can create a brand new undertaking and begin afresh.

When loading CSVs that comprise date columns, I attempt to use the parse_dates argument in pd.read_csv(). This protects me the step of manually changing the column sort later, making certain the date options are right from the beginning.

import pandas as pd
df = pd.read_csv('online_retail_clean.csv', parse_dates=['InvoiceDate'])
print(f"Efficiently loaded clear knowledge. Complete rows: {len(df)}")

Output:

Efficiently loaded clear knowledge. Complete rows: 52933

Product Deep Dive: Who Are the True Cash Makers?

With the clear knowledge loaded, let’s bounce straight into product efficiency. Merchandise could be thought of “best-sellers” in two alternative ways: by quantity (items offered) or by worth (income generated).

Perception 2: Prime 10 Bestsellers (by Amount)

The amount checklist tells us which merchandise are standard and in excessive demand, even when their value is low. That is necessary for warehouse stocking and stock administration.

We’ll group by the product Description, sum the Amount, and kind it descendingly.

# Prime 10 Bestsellers (by Amount)
df.groupby(‘Description’)[‘Quantity’].sum().sort_values(ascending=False).head(10)

Output:

Description
MEDIUM CERAMIC TOP STORAGE JAR 74698
WORLD WAR 2 GLIDERS ASSTD DESIGNS 9419
GIN + TONIC DIET METAL SIGN 5016
ASSORTED COLOURS SILK FAN 4684
JUMBO BAG RED RETROSPOT 4671
PACK OF 72 RETROSPOT CAKE CASES 3561
MINI PAINT SET VINTAGE 3285
ASSORTED COLOUR BIRD ORNAMENT 2783
BROCADE RING PURSE 2525
POPCORN HOLDER 2508
Identify: Amount, dtype: int64

Perception 3: Prime 10 Income Drivers (The True Cash-Makers)

That is what NovaShop actually needs to know. These merchandise drive profitability. They may not promote as many items because the gadgets within the earlier checklist, however their increased value or bigger bulk orders make them much more precious.

We observe the identical course of, however this time we combination on the Income column we engineered in Half 1.

# Prime 10 Income Drivers (The True Cash-Makers)
df.groupby(‘Description’)[‘Revenue’].sum().sort_values(ascending=False).head(10)

Output:

Description
MEDIUM CERAMIC TOP STORAGE JAR 77785.74
DOTCOM POSTAGE 18456.13
REGENCY CAKESTAND 3 TIER 15093.38
AMAZON FEE 13541.33
PARTY BUNTING 9491.60
GIN + TONIC DIET METAL SIGN 9480.51
JUMBO BAG RED RETROSPOT 8996.65
Guide 8966.96
CHILLI LIGHTS 8395.20
POSTAGE 7194.51
Identify: Income, dtype: float64

Dialogue: What These Lists Inform NovaShop

That is the place the ability of EDA really shines, as a result of these two lists inform a really totally different story:

• The All-Star Product: The “MEDIUM CERAMIC TOP STORAGE JAR” is clearly the winner, dominating each lists by an enormous margin. It’s excessive quantity and excessive worth. This merchandise needs to be NovaShop’s main focus for stock, advertising, and growth.
• A shocking income supply: Whereas exploring the info, I seen gadgets like “DOTCOM POSTAGE,” “AMAZON FEE,” “Guide,” and “POSTAGE.” These aren’t precise merchandise — they’re service charges and handbook changes. But collectively, they create in nearly £48,000 in income. NovaShop ought to double-check whether or not these charges are being counted of their gross revenue. It’s a straightforward factor to overlook, nevertheless it might be a quiet and significant supply of earnings hiding in plain sight.
• The Excessive-Worth Bodily Items: Gadgets just like the “REGENCY CAKESTAND 3 TIER” and “CHILLI LIGHTS” make the income checklist however not the amount checklist. This tells us they’re high-price gadgets that contribute considerably to the overall money move, and they’re price focused advertising efforts.

We now have a whole image of what NovaShop sells greatest. Time to determine when!

Making ready for Time-Sequence Evaluation: Function Engineering

We all know what offered nicely. Now, let’s determine when gross sales peak. Our purpose is to interrupt down the InvoiceDate column into parts that enable us to group and analyze gross sales by yr, month, day, and even hour.

This can be a basic characteristic engineering transfer. Since we already made certain InvoiceDate is a correct Pandas datetime object (due to parse_dates in our loading step!), we will use the extraordinarily helpful .dt accessor.

Extracting Time Elements

To know seasonality (month-to-month) and operational effectivity (hourly), we’ll create 4 new columns.

• 12 months — For long-term comparisons
• Month — For figuring out seasonal peaks (e.g., This fall rush).
• DayName — To seek out the busiest day of the week.
• Hour — To find out peak operational hours.

Let’s generate these columns:

print(“n — — Time Sequence Function Extraction — -”)

# Extract 12 months, Month, DayName, and Hour from the InvoiceDate column
df[‘Year’] = df[‘InvoiceDate’].dt.yr
df[‘Month’] = df[‘InvoiceDate’].dt.month
df[‘DayName’] = df[‘InvoiceDate’].dt.day_name()
df[‘Hour’] = df[‘InvoiceDate’].dt.hour
print(“New time options added: 12 months, Month, DayName, Hour.”)
print(df[[‘InvoiceDate’, ‘Year’, ‘Month’, ‘DayName’, ‘Hour’]].head())

Output:

--- Time Sequence Function Extraction ---
New time options added: 12 months, Month, DayName, Hour.
InvoiceDate 12 months Month DayName Hour
0 2011-06-01 12:05:00 2011 6 Wednesday 12
1 2011-05-27 17:14:00 2011 5 Friday 17
2 2011-04-21 17:05:00 2011 4 Thursday 17
3 2011-11-16 10:39:00 2011 11 Wednesday 10
4 2011-03-10 08:40:00 2011 3 Thursday 8

Good! We are able to transfer proper into mapping the gross sales patterns within the subsequent part.

Mapping Gross sales Patterns (Time-Sequence Insights)

With our time options prepared, we will now ask and reply our questions on when gross sales happen. We’ll begin large (month-to-month traits) and drill down small (hourly traits).

Perception 4: Gross sales Traits by Month (Seasonality)

Month-to-month aggregation is essential for recognizing seasonality — the predictable sample of gross sales repeating yearly. This knowledge is used for monetary forecasting and planning stock ranges for the height season.

We’ll group by the Month column and calculate the overall income for every.

print(“n — — Perception 4: Gross sales Traits by Month (Seasonality) — -”)
# Group by month and sum the income
monthly_revenue = df.groupby(‘Month’)[‘Revenue’].sum().sort_values(ascending=False)
# Print the outcomes, which ought to present the seasonal peaks
print(monthly_revenue)

Output:

--- Perception 4: Gross sales Traits by Month (Seasonality) ---
Month
11 143576.360
1 142013.420
12 133178.980
10 119533.540
9 103929.420
3 72968.270
8 71302.890
5 70424.510
6 68722.520
7 68713.831
4 51882.010
2 50178.730
Identify: Income, dtype: float64

Interpretation: The Huge This fall and January Shock

The information clearly reveals that NovaShop’s gross sales are closely skewed towards the top of the yr, confirming the standard retail This fall (October, November, December) rush.

• Peak Season: November is the height month by a slight margin, adopted intently by December. That is the time to replenish on the Medium Ceramic Prime Storage Jars and run focused adverts.
• The January Surge: The second-highest month is January! This is a wonderful perception for NovaShop. It suggests clients could also be utilizing Christmas reward cash or benefiting from post-holiday gross sales. This era shouldn’t be handled as a stoop however as a secondary high-volume gross sales alternative.
• Low Season: February and April are the bottom performers, which helps NovaShop plan for decrease staffing and advertising budgets throughout these durations.

Perception 5: Gross sales Traits by Day of Week and Hour (Operational Planning)

Whereas seasonality is about finance and stock, every day and hourly traits are about operations. NovaShop can use this to schedule warehouse employees, optimize advert spend, and time e mail campaigns.
We’ll run two separate aggregations right here: one for the day of the week and one for the hour of the day.

Income by Day of Week

# 1. Gross sales by Day of Week
daily_revenue = df.groupby(‘DayName’)[‘Revenue’].sum()
print(“n — — Income by Day of Week — -”)
print(daily_revenue)

Output:

--- Income by Day of Week ---
DayName
Friday 163861.320
Monday 171026.230
Sunday 83125.890
Thursday 219342.980
Tuesday 282796.741
Wednesday 176271.320
Identify: Income, dtype: float64

Interpretation: Tuesday is the Energy Day

The information reveals a transparent and actionable sample for the work week:

• Peak Gross sales: Tuesday completely dominates the week, producing nearly 30% extra income than the following busiest day, Thursday. This can be a prime day for launching new merchandise or working high-impact flash gross sales.
• Hunch Days: Unsurprisingly, gross sales drop dramatically on Sunday. NovaShop would possibly take into account this a low-priority day for e mail advertising or customer support staffing.
• Lacking Day: The dataset comprises no transactions for Saturday in our pattern (or maybe the complete set, which is widespread in B2B datasets).

Income by Hour of Day

# 2. Gross sales by Hour of Day
hourly_revenue = df.groupby(‘Hour’)[‘Revenue’].sum().sort_index()
print(“n — — Income by Hour of Day — -”)
print(hourly_revenue)

Output:

--- Income by Hour of Day ---
Hour
7 2830.910
8 26597.800
9 76768.750
10 209809.971
11 115233.600
12 142474.460
13 130348.290
14 119241.520
15 145178.980
16 70620.550
17 36148.030
18 14022.790
19 4397.130
20 2751.700

Interpretation: The Late Morning Growth

The hourly knowledge factors on to operational wants:

• Huge Peak: Gross sales explode at 10 AM (Hour 10), producing nearly double the income of the following few hours. That is the optimum time for NovaShop to launch e mail advertising and social media campaigns.
• Sustained Exercise: Gross sales stay excessive from 11 AM via 3 PM, suggesting a powerful noon and post-lunch buying window.
• Finish of Day: Exercise falls off sharply after 4 PM, with little or no income generated within the night (7 PM onward). This means NovaShop’s buyer base is primarily desk-based, buying throughout conventional working hours.

Conclusion: Setting the Stage for Buyer Segmentation

We’ve formally completed the heavy-duty EDA! We all know what sells greatest (the Ceramic Jar and costs), the place gross sales are strongest (the UK), and when the gross sales peaks happen (This fall, Tuesdays, and 10 AM).

Nevertheless, we’re nonetheless lacking probably the most precious piece of the puzzle: The Who. The very best firms don’t simply know what sells; they know who their most beneficial clients are and the best way to hold them.

 In Half 3, we’ll dive into probably the most superior evaluation approach on this collection: RFM (Recency, Frequency, Financial) Buyer Segmentation. This may enable us to categorise clients into tiers like “Champions,” “Loyal Clients,” and “At-Danger,” giving NovaShop a real roadmap for personalised advertising.

I hope you discovered this text useful. Be at liberty to say hello to me on any of those platforms. I really recognize your suggestions.

Medium

LinkedIn

Twitter

YouTube