DATA SCIENCE SERIES

Univariate, Bi-variate, and Multivariate analysis using IBM SPSS Modeler

Howdy! 👋
I’m a UX designer, who got interested in Business analysis using data mining techniques. I hope you find this article insightful 😉

It was inspired by this amazing article by Irfan Fadhullah

Dataset 🛒

Find this data set on Kaggle 🤘

Dataset context

The growth of supermarkets in most populated cities are increasing and market competitions are also high. The dataset is one of the historical sales of supermarket company which has recorded in 3 different branches for 3 months data.

Attribute information

1. Invoice id: Computer generated identification number
2. Branch: Branch of supermarket (A, B, C)
3. City: Location of supermarket (Yangon, Naypyitaw, Mandalay)
4. Customer type: Type of customers (Normal, Member)
5. Gender: Gender of the customers (Male, Female)
6. Product line: Product categories (Food and beverages, Health and beauty, electronic accessories, Fashion accessories, Home and lifestyle, Sports and travel)
7. Unit price: Price of products in US dollars
8. Quantity: Number of products purchased by each customer
9. Tax: 5% tax fee on top of purchase amount
10. Total: Total price of the purchase including tax
11. Date: Date of purchase (January 2019 to March 2019)
12. Time: Time of purchase (10am to 9pm)
13. Payment: Payment methods (E-wallet, Cash, Credit card)
14. COGS: Cost of Goods Sold
15. Gross margin percentage: Gross margin percentage of the purchase
16. Rating: Customer stratification rating from each customer (From 1 to 10)

Business understanding 🕴

Client

Supermarket chain in Myanmar

Business goal

To improve the gross income of the supermarket C (Naypyitaw)

Business questions

1. Segment the “high profit” and “low profit” customers to devise a customized marketing strategy for boosting the sales of supermarket C.
2. How’s gross income affected by unit price, quantity, and other variables like day, timeslot, and product line in general?
3. How’s the unit price affected by gross income, quantity, and other variables like day, timeslot, and product line?
4. Find and explain the relationship between gender and product line, payment, gross income for Branch C?
5. Find and explain the relationship between customer type and gender, day, timeslot for Branch C?
6. Predict the day of future purchases of the customers.

Data mining goal

1. Create a clustering model to segment the high profit and low profit customers of supermarket C.
2. Create a model to analyze the relationship between different variables of the supermarket.
3. Create a prediction model to predict the day of future purchase.

Data understanding 🤔

Field type analysis

It is evident that Invoice ID is detected as “typeless” since it has hyphens in between which could not be detected as a valid measurement type. Whereas, fields like branch, city, product line and payment are “nominal” that have multiple values. The customer type and gender are a binary field, and thus has “flag” measurement. And the rest of the variables are “continuous” since they have a range of numeric values.

Record count

Let’s go through briefly to analyze the number of records and individual field count to understand the overall state of the dataset. I’ve used aggregate and table nodes to find the specific record count.

This dataset contains 1000 records with individual branch distribution being 340 Branch A records, 328 Branch B records, and 332 Branch C records. The customer type and gender are equally distributed with 501 Member and 499 Normal customer records, 499 Male and 501 Female records. The product line has a wide distribution with 170 Electronic accessories, 178 Fashion accessories, 174 Food and beverages, 152 Health and beauty, 160 Home and lifestyle, and 166 Sports and travel records. The payment method is distributed into 3 categories with 345 e-wallet, 344 cash, and 311 credit card records.

Uni-variate analysis 🧐

3.3.1 Is there any difference in aggregate sales across all the branches?

A distribution graph node was used to find the frequency of the categorical field “Branch”. There is no much of a difference in sales across all the branches A,B, and C. But, the sales in Branch A is a bit higher than the rest of the branches. Here’s the numeric data of the branch sales,

a. Branch A — 340

b. Branch B — 332

c. Branch C — 328

3.3.2 Which is the most popular payment method that is used by the customers?

A distribution graph node was used to find the frequency of the categorical field “Payment”. The E-wallet payment method is the most popular one, but cash is also very popular. The least used payment method is the credit card. Here’s the numeric data of the payment method used,

a. Cash — 344

b. Credit card — 311

c. Ewallet — 345

3.3.3 Which is the most popular product line across all the branches?

A distribution graph node was used to find the frequency of the categorical field “Product line”. The most popular product line is the Fashion accessories, and the second most popular product line is the Food and beverages. Here’s the numeric data of different product lines (purchase counts),

a. Electronic accessories — 170 b. Fashion accessories — 178 c. Food and beverages — 174 d. Health and beauty — 152 e. Home and lifestyle — 160 f. Sports and travel — 166

3.3.4 Which gender makes the greater number of purchases?

A distribution graph node was used to find the frequency of the flag field “Gender”. It is evident that people from both the genders are making almost same number of purchases. But females tend to make more than male when we look more precisely. Here’s the numeric data of purchases made based on gender data,

a. Male — 499

b. Female — 501

Bi-variate analysis 🤓

3.4.1 Does gross income affect the ratings that the customers provide?

A scatter plot node was used to generate a scatter plot with a trend line between gross income and rating. Since the trend line is quite flat, we can conclude that gross income does not have a significant effect on customer stratification rating.

3.4.2 Which customer type is more profitable when customer stratification rating is considered?

The plot node was used to visualize the data in the form of stacked histogram. It is visually evident that customer who belong to Member category purchase more than the Normal customers. Thus, Members provide higher gross income compared to normal customers.

And, the customers who belong to customer stratification rating range of 6 to 8 purchase more than the other rating ranges.

Here’s the numeric data of gross income from both the customer types,

Member = $7820.164

Normal = $7559.205

Here’s the numeric data of gross income from customers who are rated between 6 to 8,

Gross income = $5564.311

3.4.3 Which is the most profitable branch?

The graphboard node was used to plot the graph between branch and gross income. It is visually evident that the Branch C is the most profitable branch among all the three. Whereas, Branch A and B are equal in profits. Here’s the numeric data of gross income from all the branches,

a. Branch A = $5,057.2

b. Branch B = $5,057.2

c. Branch C = $5,265.2

3.4.4 Which gender brings more profit to the supermarket?

The graphboard node was used to visualize this graph between gender and gross income. It is visually evident that Females majorly contribute to the supermarket’s profit. Here’s the numeric data for gender-wise gross income,

a. Male = $7,384.9

b. Female = $7,994.4

3.4.5 Which is the most profitable product line?

The graphboard node was used to plot the graph between gross income and product line. It is evident that the Food and beverages category is the most profitable one, and the Sports and travel is the second most profitable one. Here’s the numeric gross income data according to each product line,

a. Electronic accessories — $2,587.5 b. Fashion accessories — $2,586 c. Food and beverages — $2,673.6 d. Health and beauty — $2,342.6 e. Home and lifestyle — $2,564.9 f. Sports and travel — $2,624.9

Correlation analysis 📉

3.5.1 How does other fields affect the gross income of the supermarket?

Statistics and graphboard node were used to find the correlation between these fields. It is very evident the gross income is strongly correlated with unit price, quantity, and total bill amount. An interesting insight is the customer stratification rating is not correlated with gross income. So, we can conclude the customers are not rated based on the profits they contribute to the supermarket or any other fields.

Data preparation 👩‍🍳

Outlier detection

No outliers were detected while using the Data Audit node.

Data cleaning

4.2.1 Remove redundant field: Tax 5%, City

As shown in the table, both of these fields are redundant. So, I decided to filter out “Tax 5%” field while using the dataset.

Since city and branch gives us almost the same information, so I decided to filter out “city” field while using the dataset.

4.2.2 Fixed the date formatting

As shown in the table, the date is not formatted in a single date style (either the US or UK). And, the separator is also mixed (hyphen and slash). So, I used “Text to columns” option in Excel to fix the date formatting issue and the separator style.

4.2.3 Create new field: day

As shown in the table, a new field called ‘Day’ was created to have the day of the respective date for generating better insights from the dataset. The ‘TEXT’ function in excel was used to convert dates into days.

4.2.4 Create new field: timeslot

As shown in the table, a new field called “Timeslot” was created to have the timeslot (morning, afternoon, evening, and night) for generating better insights from the dataset. Using the ‘Derive’ node and set the multiple nominal rules to segment the data into 4 different categories.

4.2.5 Removal of unwanted field: Invoice ID, Gross margin percentage

Just to have the required fields for further data modelling, I decided to filter the Invoice ID field. Also, the Invoice ID has the measurement ‘Typeless’ which does not add any significant impact on further steps.

Since the gross margin percentage is the same ‘4.762%’ it does not have any impact for my business questions.

Additional analysis 📊

After I created new variables, I would like to analyze and know some answers for some of the questions. This can provide a different perspective towards the business goal.

5.1 Which day of the week is highly profitable in gross income?

The graphboard node was used to plot the bar chart of gross income vs day. It is very evident that the highest profitable day during the week is ‘Saturday’ followed by ‘Tuesday’. And, Sunday, Thursday, and Wednesday performs almost similarly. Here’s the numeric data of gross income per each day,

a. Monday = $1,804.7

b. Tuesday =$2,451.5

c. Wednesday =$2,082.4

d. Thursday = $2,159.5

e. Friday = $2,091.7

f. Saturday = $2,672.4

g. Sunday = $2,117.0

5.2 Which timeslot of the day is most profitable in gross income?

The graphboard node was used to plot the bar chart between gross income and timeslot. It is evident the profits are higher during the afternoons, followed by morning timeslot. Here’s the numeric data of gross income per each timeslot,

a. Morning = $2,942.8

b. Afternoon = $7,048.7

c. Evening = $2,4414.

d. Night = $2,946.4

5.3 Which day is highly profitable in gross income in the morning, afternoon, evening, and night?

The graphboard node was used to visualize the box plot of gross income vs day & time slot. Below is the detailed numeric data of gross income,

During the mornings,

1. Sunday = $19.057
2. Wednesday = $17.711

During the afternoon,

1. Tuesday = $18.054

During the evening,

1. Saturday = $18.744

During the night,

1. Thursday = $18.358
2. Friday = $18.283

Thus, supermarket can take advantage of this trend and provide the best experience to its customers. Here’s the numeric

Modelling 📳

Clustering

5.1.1 Segment the “high profit” and “low profit” customers to devise a customized marketing strategy for supermarket C along with additional characteristics.

Auto Cluster node was used to quickly find the best cluster having the best silhouette value using 2 different clustering methods (K-means and two-step).

Cluster analysis helps to segment different customers with their unique characteristics that can help the supermarket to devise a customized marketing strategy.

K-means & two step clustering

A select node was used to select the data only from branch C, and a partition node was used to split the data into 80–20 to avoid the over-fitting problems. Since the data has not labeled the customers into high profit and low profit, also the customer stratification rating is not directly related to the gross income, we have to cluster the dataset based on the following features,

1. Customer type
2. Product line
3. Day
4. Timeslot
5. Payment
6. Gross income

Considering the best silhouette value, I plotted the graph between number of clusters and their respective silhouette value to select the best model,

After analyzing the silhouette value and the cluster quality for both the clustering methods, I have decided to go for K-means clustering consisting 6 clusters with the second highest silhouette value of 0.338.

As we can see the cluster quality is almost fair with no potential overfitting. Now, let us see all the 6 clusters formed using the K-means algorithm.

Describing the high profit cluster,

1. Size: It is the second largest cluster (18.5%) among all the 6 clusters
2. Preferred payment: Ewallet by everyone (100%)
3. Customer type: Normal (100%)
4. Gross income: $17.56 (higher than the median of $12.925)
5. Frequent day: Thursday (26.1%) followed by Tuesday and Saturday
6. Frequent timeslot: Afternoon (60.9%)
7. Frequently purchased product line: Fashion accessories (23.9%) followed by Food and beverages, Health and beauty

Inference: It is surprising that the normal customers are generating the highest profits (gross income) when compared to member customers. They buy mostly on Thursday afternoons using their E-wallet for payments. Also, they mostly purchase products from Fashion accessories.

Describing the low profit cluster,

1. Size: It is the smallest cluster (11.6%) among all the 6 clusters
2. Preferred payment: Credit card by everyone (100%)
3. Customer type: Normal (100%)
4. Gross income: $12.28 (lower than the median of $12.925)
5. Frequent day: Wednesday (24.1%) followed by Tuesday, Thursday, and Saturday
6. Frequent timeslot: Afternoon (44.8%)
7. Frequently purchased product line: Fashion accessories (24.1%) followed by Health and beauty

Inference: It is evident that the normal customers belong to this cluster with the least gross income of $12.28. They mostly purchase during the Wednesday afternoons with their credit card to buy products from fashion accessories.

Linear regression

Linear regression was used to find linear relationships between different continuous variables. I tried to have multiple variations with different target variables and different splits, but no significant difference in the R2 score was found.

5.2.1 How’s gross income affected by unit price, quantity, and other variables like, day, timeslot, and product line.

Since the question also asks about the variables which are nominal, I had to convert them to flag based variables so that I could use them in the linear regression modelling. Here are the variables involved in the modeling,

Input variables,

1. Unit price — Continuous
2. Quantity — Continuous
3. Day — Nominal → Flag
4. Timeslot — Nominal → Flag
5. Product line — Nominal → Flag

Target,

1. Gross income — Continuous

The SetToFlag node was used to convert the nominal to flag,

It is evident that the R2 score remains the same even if the partition split changes,

We can conclude the R2 score being substantial (0.892), and the data fit the model quite well. Also, the statistical significance is 0.000 so the results are significant.

Here’s the findings from regression model,

It is very conclusive from the coefficients that quantity (2.796) being the most impactful variable on the gross income of the supermarket. Also, the unit price (0.2772) with an interesting insight of Tuesday’s sales (0.2211) with payment made via credit card (0.519) has also had a positive impact. Finally, the Home and lifestyle products (0.3098) contribute the most and positively to the supermarket’s gross income.

5.2.2 How’s rating affected by unit price, quantity, and gross income?

It was evident with the R2 score of 0.004 is the lowest and is bad for any conclusions to draw from the model.

5.2.3 How’s the unit price affected by gross income, quantity, and other variables like day, timeslot, and product line.

Since the question also asks about the variables which are nominal, I had to convert them to flag based variables so that I could use them in the linear regression modelling. Here are the variables involved in the modeling,

Input variables,

1. Gross income — Continuous
2. Quantity — Continuous
3. Day — Nominal → Flag
4. Timeslot — Nominal → Flag
5. Product line — Nominal → Flag

Target,

1. Unit price — Continuous

The SetToFlag node was used to convert the nominal to flag,

It is evident that the R2 score remains the same even if the partition split changes,

We can conclude the R2 score being substantial (0.785), and the data fits the model quite well. Also, the statistical significance is 0.000 so the results are significant.

Here’s the findings from regression model,

It is conclusive that the gross income (2.823) has the highest impact on the unit price. Also, Monday (2.372), Morning (1.767), Cash payment (0.1373) has a positive relationship with the unit price.

Logistic regression

Logistic regression was used to find the relationship between variables and the impact of one on the other variable. In general, I tried to answer the below mentioned in 4 different ways to find the best accuracy. Here are the four combinations of training & testing the dataset to achieve the best accuracy,

1. Full dataset with 80–20 split
2. Full dataset with 75–25 split
3. Dataset of only Branch C with 80–20 split
4. Dataset of only Branch C with 75–25 split

Here’s the Full dataset version,

Here’s the Branch C version,

5.3.1 Find and explain the relationship between gender and product line, payment, gross income for Branch C.

Let’s first see the accuracies of all the four versions,

a. Full dataset with 80–20 split — 52.4%

b. Full dataset with 75–25 split — 55.82%

c. Dataset of only Branch C with 80–20 split — 41.77%

d. Dataset of only Branch C with 75–25 split — 41.3%

As we know from the previous data, females contribute more towards supermarket’s gross income than males. So, I decided to set the base category for target as females, and get more insights on male.

Some of the insights from the regression equation are,

1. (-0.1909) Male customers contribute the least to gross income while spending with cash
2. (0.8547) Male customers also contribute more to gross income while spending on Home and lifestyle products
3. (0.1328) Male customers mostly purchase Home and lifestyle products globally, and Fashion accessories paid via cash
4. (-0.8603) Similarly, electronic accessories are least or never purchased by male customers via cash
5. (-0.2663) Male customers do not purchase Home and lifestyle products via credit card

5.3.2 Find and explain the relationship between customer type and gender, day, timeslot for Branch C.

Let’s first see the accuracies of all the four versions,

a. Full dataset with 80–20 split — 46.77%

b. Full dataset with 75–25 split — 51.81%

c. Dataset of only Branch C with 80–20 split — 51.9%

d. Dataset of only Branch C with 75–25 split — 53.26%

Finding the potential normal customers to convert them to members could be beneficial to the supermarket, and thus I have set the members as base category.

Some of the insights from the regression equation are,

1. (0.5286) Normal customers do their shopping on Monday are the highest
2. (0.4463) They prefer mornings to do their shopping
3. (0.7156) Normal customers who are female and shop during Thursday are the highest
4. (1.673) Normal customers who are female and shop during Tuesday afternoon are the highest
5. (-1.413) Normal customers who shop on Monday mornings are the least
6. (-0.7133) Normal customers who are female and shop on Friday are the least

5.3.3 Find and explain the relationship product line between branch, customer type, gender, day for Branch C.

Let’s first see the accuracies of all the four versions,

a. Full dataset with 80–20 split — 17.91%

b. Full dataset with 75–25 split — 16.47%

c. Dataset of only Branch C with 80–20 split — 12.66%

d. Dataset of only Branch C with 75–25 split — 10.87%

Since the Health and beauty are the worst performing product line in terms of gross income from the previous data, I decided to focus on that.

Some of the insights from the regression equation are,

1. (0.2857) Health and beauty product line perform better at Branch B
2. (0.4387) Health and beauty product line perform better on Friday
3. (0.7856) Health and beauty product line perform better on Sunday in Branch A
4. (1.714) Health and beauty product line perform better on Friday in Branch B
5. (1.191) Health and beauty product line perform better on Sunday’s with Member customers
6. (-1.828) Health and beauty product line perform the worst on Sunday’s by Female

To summarize, the 75–25 split had better accuracy rates but with slightly different input datasets. But, each question is answered with higher accuracy with different data combinations.

Classification: Nearest Neighbor

Due to pre-classified data of product lines and fewer continuous inputs I wasn’t able to find any conclusions or insights. Also, this was not a classification problem. So, I found this modeling method not applicable to my dataset.

Prediction: Neural networks & Random forests

I chose two algorithms to predict the day of purchase while taking some of the features from the dataset. To get a different perspective, I converted the product line nominal variable to a continuous number-based variable. Here’s how I converted the nominal to continuous,

First, I selected the field I wanted to change and assigned new values to each option.

And, since these new values were in string format, I had to change them to number format.

Now, let’s explore the inputs and target used for training both the models,

Inputs,

1. Quantity — Continuous
2. Unit price — Continuous
3. Total — Continuous
4. Product line (R) — Continuous

Target,

1. Day — Nominal

Let’s now see how both the models perform with the same inputs and target of the same dataset,

Neural networks

The model did not perform really well. It was not able to detect the days accurately when compared to the actual data. This can be concluded with a poor Correct % of just 13.41%.

Random Forest

The model performed really well being able to predict very the days very accurately when compared with the real data. The Correct % was 96.34%.

To conclude, Random Forest was the best model to predict the day of purchase in the given dataset.

Prediction: Decision trees (C 5.0, CHAID)

I chose two algorithms to predict the day of purchase while taking some of the features from the dataset. To get a different perspective, I converted the product line nominal variable to a continuous number-based variable. Here’s how I converted the nominal to continuous,

First, I selected the field I wanted to change and assigned new values to each option.

And, since these new values were in string format, I had to change them to number format.

Now, let’s explore the inputs and target used for training both the models,

Inputs,

1. Quantity — Continuous
2. Unit price — Continuous
3. Total — Continuous
4. Product line (R) — Continuous

Target,

1. Day — Nominal

Let’s now see how both the models perform with the same inputs and target of the same dataset,

C 5.0 & CHAID node was used to predict the day of purchase

I’ve used both models in a direct fashion to get the individual prediction. Let’s compare the correct percentage from each model and different data split percentage,

C 5.0 (80–20 split) The correct percentage & equation,

CHAID (80–20 split) The correct percentage & equation,

C 5.0 (75–25 split) The correct percentage & equation,

CHAID (75–25 split) The correct percentage & equation,

Until now, it is conclusive that these models are not better than the random forest model. Let’s try a different combination with the introduction of logistic regression node with the same split percentages,

C 5.0 + CHAID + Logistic regression (80–20 split),

C 5.0 + CHAID + Logistic regression (75–25 split),

So, we can conclude again that random forest was the best model for predicting the day of purchase.

Evaluation 🙋‍♂️

Question 1: Segment the “high profit” and “low profit” customers to devise a customized marketing strategy for boosting the sales of supermarket C.

Results: The low-profit customer cluster prefers credit card payments, and everyone is normal customers. They mostly purchase on Wednesday afternoons and mostly buy Fashion accessories.

Business action: Provide cashback and discount offers for purchases made via credit cards, and promote the supermarket memberships to the normal customers.

Question 2: How’s gross income affected by unit price, quantity, and other variables like day, timeslot, and product line in general?

Results: It is very conclusive from the coefficients that quantity (2.796) being the most impactful variable on the gross income of the supermarket. Also, the unit price (0.2772) with an interesting insight of Tuesday’s sales (0.2211) with payment made via credit card (0.519) has also had a positive impact. Finally, the Home and lifestyle products (0.3098) contribute the most and positively to the supermarket’s gross income.

Business action: Promoting the sales of Home and lifestyle products through multiple communication channels, and exclusive Tuesday sales can help the supermarket improve its profits. Finally, selling high-priced products can really boost gross income.

Question 3: How’s the unit price affected by gross income, quantity, and other variables like, day, timeslot, and product line?

Results: It is conclusive that the gross income (2.823) has the highest impact on the unit price. Also, Monday (2.372), Morning (1.767), Cash payment (0.1373) has a positive relationship with the unit price.

Business action: Having cross-selling campaigns and exclusive combo offers for monday mornings on high priced products to boost the gross income of the supermarket.

Question 4: Find and explain the relationship between gender and product line, payment, gross income for Branch C.

**Results: (**0.1909) Male customers contribute the least to gross income while spending with cash. (0.8547) Male customers also contribute more to gross income while spending on Home and lifestyle products. (0.1328) Male customers mostly purchase Home and lifestyle products globally, and Fashion accessories paid via cash (0.8603). Similarly, electronic accessories are least or never purchased by male customers via cash (0.2663). Male customers do not purchase Home and lifestyle products via credit card.

Business action: Promoting the offers and discounts on male products, and bundling them with top-selling female products to increase the overall profits

Question 5 & 6: Find and explain the relationship between customer type and gender, day, timeslot for Branch C? Predict the day of future purchases of the customers.

**Results: (**0.5286) Normal customers who do their shopping on Monday are the highest. (0.4463) They prefer mornings to do their shopping (0.7156) Normal customers who are female and shop during Thursday are the highest. (1.673) Normal customers who are female and shop during Tuesday afternoon are the highest. (1.413) Normal customers who shop on Monday mornings are the least. (0.7133) Normal customers who are female and shop on Friday are the least.

Business action: Having exclusive member discounts on Monday, Thursday, and Tuesday can help increase the member subscription revenue. Plus, it will also increase product sales.

I have summarized my whole process, and I hope it helped you!

Buy me a beer 🍺

You can support me to write more of such articles here

You can clap 50 times 👏, did you know?

Click and hold the claps icon until you reach 50

Hope this helps you to gear-up! Feel free to reach out to me on,
👨‍💼 LinkedIn | 💌 Email | 🐤 Twitter

Uncovering the most common problems faced in the 21st century

This article provides deeper insights into some of the problems faced by the internet and tech users of this era and some tested methods to combat those.

Internet, the most helpful and evolving tool for mankind. Some of the most impeccable inventions 🚀 and discoveries are possible today because of the Internet.

But, have you ever noticed your focus being drastically reduced while reading a lengthy article (Don’t worry, I’ve tried to keep this short & useful😛), or a chapter from a book? Have you ever faced difficulty in recalling some important information that was perceived very recently? And, quickly reached out to your phone to fetch the same? 🙋‍♂️

Chill out, We’ve all had a history of remembering stuff really well especially during our school exams (They completely su*k). I specifically remember my schooling in India, where I had to memorize 2 to 3 pages long essays and question answers while preparing for my final exams. Most of my studies happened in books, and I used to love reading books. But, I don’t feel like reading books or reading lengthy articles anymore, I’d rather watch Game of Thrones or a nice movie 😎

If you love reading books, pls ignore me, I’m lying 😂

We are getting used to quickly check our notifications, messages, and take further action on them by replying back to a friend or take a quick glance of an article. Most of the apps over there are designed by highly skilled UX Designers and Copywriters for retaining users within the app and bring them back to the app. This behavior rewires our brain for skimming short snippets of information at once and in a regular time-interval. For example, notifications from social media apps, a new message, or an email alert. Gradually, we are aiming for information that can be consumed quickly and we are learning the habit of procrastination if we encounter lengthy content.

In addition, there are so many interesting resources available today for information consumption which makes us overwhelmed before even starting with something. Some common behaviors are 10–15+ tabs opened on our browser, unable to finish reading a single webpage. This is mainly because the web psychology & UX is designed to redirect users to new links and webpages as they start skimming a webpage — This eventually results in multiple tabs + Incomplete content + Excessive overwhelmingness.

The number of tweets, likes, posts is exponentially increasing, but the distance between our own kind is also exponentially increasing as well. It almost took this Pandemic to stay home and spend some quality time with our family and loved ones, but still, we are trying to keep ourselves engaged with some or the other thing. We all have 1000’s of friends on our social platforms but, we struggle to share any emotional matter or information with any of those social friends. I think Social media has formalized our connections with our friends and family. Of course, I agree that the internet has helped us to build strong communities across the world, but we are drifting away from people who are physically close to us 😥

I know all this might sound lame when we all can see some great discoveries and breakthroughs made in the history of mankind using tech. But, we all envision the amazing to happen to our future and stop to live in present and enjoy the moment. I also agree with our competitive and fast-moving world which comes with an additional gift like stress, anxiety, depression, and more… I deliberately remembered a popular Tamil saying, “வெளுத்ததெல்லாம் பாலல்ல மின்னுவதெல்லாம் பொன்னல்ல” which means Everything which is white in color is not as safe as milk, and everything which glows is not as pure as gold. Our online society can seem very sweet and safe for us, but it isn’t.

As I mentioned earlier, internet communities are growing rapidly and the platform to share thoughts is also increasing. If everyone is engaged in sharing their thoughts, who’s left to listen and make calculated decisions? I know that it’s our fundamental right to express our thoughts, but we are losing our patience, the capability to listen, and becoming quick-witted which might not help us in the long run. As we grow in age as well as in career listening patiently comes very handy, but with an extra filter to it to bring down incoming noise or garbage data to zero. I’ve seen many people who boldly express their opinions and advice online but, unable to do it with real people. It’s always very easy to share and comment across social platforms, but it's hard to follow it as well as to convey the same information to a real person patiently. Attention to all the designers, Listening is one of the most asked soft skills from companies and Hiring Managers.

It takes time to practice it though 🙈

Being a UX Designer, it’s my job to make quick and well-thought (Already contradicting 😂) design decisions while crafting an interface. It’s our need to save time by quickly checking our notifications, quickly have our food and water, quickly reach office, quickly reply to our messages. But, have you ever tried to ignore your notifications and enjoy the moment? Have you ever tried to enjoy the flavor and fragrance of your food🥘? Have you ever tried to wake up early in the morning, finish your morning chores, and reach office on time without any hurry? Again, we can’t afford to sit and do everything patiently, but managing our time in a better way, making calculated decisions, worrying too much about what’s coming up next can help all of us :)

Our generation needs everything quick — internet speed, job email response, a girl’s reply for a proposal, and what not? We’ve almost lost our patience & inculcated the quality of worrying too much. With the speed comes FOMO (Fear Of Missing Out), it’s very common these days. There’s a ton of new things happening around us that are sourced by social media and news. It mostly happens with lower satisfaction in our lives which triggers us to find something happy or engaging online. This thing is going crazy, I literally found a social marketing company called Fomo using this technique to increase sales 🤣. It’s funny at the same time scary to think about how companies can manipulate our human behaviors to buy something online or spend more time on their sites. Its always good to take a step back and think mindfully.

Next comes the knight “Overthinking!” We all are very conscious of our social and online presence nowadays. When don't stop thinking and re-evaluating our actions and reactions, it can interfere with our quality of life. This is usually the result of increased anxiety (A mental state characterized by feelings of worry or panic). It happens when we mainly engage with all possible ‘what-ifs’ of a particular situation, and we end up thinking for hours and hours but still unable to come to a conclusion. When we run through a chain of thoughts, again and again, we eventually develop stress and depression which also leads to bad sleep quality. Moderate-thinking (I just made that up 🤪) sometimes helps designers to find out all the possible edge cases by running through every use case. But, this does not hurt our mental state very much because the thinking isn’t related to us but the product which we are building.

Speaking of work, I usually check if any new messages or email pops-up. If there are any quick actionables, I’ll do it immediately by transferring the relevant link to my computer. As I get away from work to finish a quick task, I get redirected into a chain of links. I will unknowingly spend like an hour or something reading, watching some other content relevant to design or technology. If we look into this, I’ve lost my attention to work and got carried away by other carefully crafted content intended to grab user’s attention. I decided to control my eagerness to check out those links by simply making a list of links to come back and check em out later. This kind of tipping point to procrastinate any of my tasks. I simply would paste the link or write the task to be done on the list and work on something else. But yeah, I’m not trying to control this thing these days by finishing small tasks instantly without getting distracted from other online content.

Some facts,

• A review by academics from Oxford, King’s College London, Harvard, and Western Sydney University, found the limitless stream of prompts and notifications from the internet encourages us towards constantly holding divided attention — which then in-turn may decrease our capacity for maintaining concentration on a single task.”
• The endless stream of notifications and digital distractions were found to be physically influencing the brain, with those affected showing less grey matter in the cerebral areas associated with maintaining focus.
• Other studies showed that the internet was also deceiving people into thinking they were smarter than they are as they “blurred the lines” between their own memories and what they can easily look up on ever-present smartphones.
• Research shows overthinkers believe they’re helping themselves by rehashing their problems in their heads. But studies show analysis paralysis is real. If you’re an overthinker, you likely already know you can’t sleep when your mind won’t shut off.
• Research conducted at Stanford University found that multitasking is less productive than doing a single thing at a time. The researchers also found that people who are regularly bombarded with several streams of electronic information cannot pay attention, recall information, or switch from one job to another as well as those who complete one task at a time.
• Researchers at the University of Sussex in the UK found that high multitaskers had less brain density in the anterior cingulate cortex, a region responsible for empathy as well as cognitive and emotional control.

Best practices to combat all the above-mentioned problems,

• Keep a 📘 notebook and a 🖊 pen to make a quick note of all the information or your thoughts while using the computer, listening to a podcast or any other event. (Digital note-making is not very efficient). This will reduce the cognitive load on your brain as well as your anxiety to remember everything and a fear of forgetting any information.
• Pomodoro technique 🍅, the technique uses a timer to break down work into intervals, traditionally 25 minutes in length, separated by short breaks (usually 3–5 minutes). After taking 4 sets of 25 min Pomodoro's, take a 15–30-minute long break.
  💻 Chrome extension for your computer: Download
  📲 Android app: Download
• Practise Mindfulness (Basically meditation) Mindfulness is the quality of being present and fully engaged with whatever we’re doing at the moment — free from distraction or judgment, and aware of our thoughts and feelings without getting caught up in them.
  🎵 Free Spotify playlist: Practise now
• Maintain a Daily Journal, I prefer Notion for it. When you start writing down all your daily learnings, interesting things you’ve explored, mistakes realized it will gradually help you to analyze and keep track of yourself.
  📕 Grab your Free Notion template: Free Template
• Maintain your action list on your phones with Google’s Tasks app. I will show you my template for making the best use of the app. I used to create different lists with labels like Pending Tasks, Daily Chores, Long-term Pending, Search queries, Work in progress. You can create your own lists and have a productive time. In addition to that, add the Tasks app widget to your home screen to be more effective.
  🎯 [Android] Download Tasks app now: Download link
  🎯 [iOS] Download Tasks app now: Download link

• Get isolated from online distractions like Ads & other marketing stuff. Use an Adblocker on your browser to stay focused.
  ✅ Free & Best Ad Blocker I’ve ever used: Download

Further reading,

• Mind-blowing facts of Internet usage: Read now
• Shorter attention spans and bad memory: Read now

Thanks for staying till the end, I’ve honestly expressed what I actually felt about my lifestyle being changed from external and internal influences…

Hope this helps you to gear-up! Feel free to reach out to me on,
🐤 Twitter | 👨‍💼 LinkedIn | 💌 Email

Just for an intro, being alone or being away from your daily schedule may trigger new thoughts, make you discover creative solutions for head-breaking problems and also will make you feel your freedom.

There are 2 types of loneliness with respect to the Time span,
a. Short time loneliness
b. Longtime loneliness

Short time loneliness is like a 100m Athletic race, which will be of a short span and you will enjoy it to the core if you are a loner.
These kinds of loneliness will make you feel refreshed from your work pressures, life problems and also remind you about yourself and your strengths! During this kind of loneliness people usually, listen to their favorite album, watch their most lovable Netflix series, sometimes go for a trip to some tourist place to enjoy and feel the nature and have some of their favorite food. Your brain will feel a neural refreshment just like clearing cache and cookies of your web browser.

Sometimes, you might also find emotional connections among your life events, people, relationships and your work. This kind of loneliness or a break will help you recover from bad transitions of your life. If this is utilized properly you might find a similar kind of happiness, as when you look into a baby’s smile and listen to her sounds! This can be utilized as a break from your social media.

And the longtime loneliness is not the same as you do in the above-mentioned scenario! It is something like a student staying back for 10 days in his University during Winter holidays when most of the students went back to their home.

This is like a marathon, testing your patience, strengths, weaknesses, and stability. This will help you know more about your deeper aspirations and goals. Become more familiar to yourself, which is the most important event of one’s life! You can achieve anything in life only if you know what you want and what to aspire to become. Longtime loneliness is similar to Software Penetration testing, which is the process of testing a computer system, network or web application to find security vulnerabilities that an attacker could exploit. This will make you undergo the most unforgettable sorrows, ups, and downs which may lead to complete depression or an enlightenment. As like a marathon, this will test each and every cell, nerve and brain power of yours in maintaining your stability until the end of the session.
Great people say you must come out of your comfort zone to achieve something big in life! Similarly, longtime loneliness process is also considered as coming out of your comfort zone, which will lead you to the solution for your biggest confusions and problems of life.

Note: Some people love this kind of loneliness, I will discuss their case in the upcoming paragraphs.

You will feel the benefits of being alone when you get back to your normal life within your comfort zone. You may feel the change and maturity attained after undergoing this process.

The Zodiac factor:

The zodiac factor also portrays its own character with the particular person, each and every zodiac sign has its own strengths and weaknesses. So there is no need to worry, be happy with what you have and make sure you know how to make the best use of your strengths and hide your weakness!

Here are some zodiac signs who hate being alone,

GEMINI (May 21 — June 20):
Gemini tops the list as one of the most social signs around. Since they love to chat, they’re always seeking out social scenes. “They enjoy meeting new people and can be energized by the social interactions they have”.

LEO (July 23 — August 22):
It’s extremely uncommon for a Leo to spend time alone, or hang out on the sidelines. Leo is often found surrounded by people. “They tend to gravitate toward the center of attention”.

LIBRA (September 23 — October 22):
Another one of the most social signs, Libra is ruled by Venus, “the planet of flirtation and social grace”, “[Their social] qualities help them establish connections with all kinds of people”.

CAPRICORN (December 22 — January 19):
Capricorn dislikes going solo, but their motivation is a bit different than the other highly-social signs. For Capricorn, they prefer engaging in social interactions that have the potential to help them get ahead in their careers.

AQUARIUS (January 20 — February 18):
Aquarius is drawn to the community and social networking, which is why they prefer going out and chatting with others over being alone. “You will find Aquarius at events that have some sort of purpose, like volunteer group outings, professional organizations, or political fundraisers”.

If you are a person who doesn’t like to be alone, here are a few suggestions:
Self-Satisfaction plays a major role by making your brain think stable and makes your body more stable.

1. Create a virtual environment of being social and try to do the same things which you used to do when you were socializing ( it’s just a try but most of them wouldn’t match this situation)
2. If you choose to listen to music never ever listen to depressing songs and the songs which you used to listen during hardships of your life. Try to remember the beautiful moments of your life.
3. Engage via social media, write blogs, create artwork and pursue any hobby which will keep you engaged. Be careful at this kind of situation, an idle mind is really a devil’s workshop!
4. If you want to make use of your lonely time for doing a productive work, try using the Pomodoro technique.
Or try watching Conjuring!!! LOL 😂😂

Remember, at this point in your life you are alone so you are not able to create something meaningful or be happy. Try to spend time doing things which you love to do and when you go deep into the work you might not feel the loneliness. At the end of the day, you will be a proud and happy person.

Results or Outcomes of your lonely time:
1. Improved self-love and self-admiration
2. You will know about the deeper insights about yourself
3. An adventure of pushing yourself out your comfort zone
4. You will come to know about your true likes and dislikes
5. You might discover your craziest desires which sometimes may change your career path as well

So, use your lonely time wisely!

While relating to S.W.Hawking, right from his Ph.D. thesis “Hawking Radiation”, “A brief history of time” to a single elegant equation which explains everything in the universe (The Theory of Everything).I was inspired by the movie “The Theory of Everything” at various scenes of it. Trying to understand various parts of it made me write this essay.

When we see the timeline of Quantum mechanics, right from Dalton’s atomic theory, JJ Thompson theory, Rutherford Alpha ray scattering experiment, Plank’s quantum theory, Bohr’s theory to Sommerfield and de Broglie’s theory, it’s been an improving theory at every stage which tells us that every theory was correct until the upgraded and contradictory version came out. It would have been correct for the person who found it, as he would live in a world of illusion of knowledge.Finally, the truth is his/her theory is wrong. A person living in his/her own illusion would never accept that his or her work would be wrong, It may be correct in his/her own perspective, but may lack a common view. The strange part of Stephen Hawking’s story is that he had moved his way towards disproving his own theory. Initially, his Ph.D. thesis stated that as time precedes the universe also grows, but if we reverse the time it may shrink and it was about “The theory of black holes and The Timespace singularity” which stated that Universe has boundaries.Then his new project disproves his Ph.D. thesis, which states that the universe has no boundaries, no beginning and there is no creator called God. When his writings were published as Hawkings radiation, which states the universe had a beginning and that made us believe that there is an existence of God. But, that was not everyone’s mistake to believe this will be the exact reason for the existence of God.

But, it wasn’t the people’s mistake but the writer’s mistake. It’s all about people who think they know something correct about a topic. And I think this is how all the great scientist’s had ended up in their research or inventions, some worked out well and some did not. Those which were irrelevant were corrected later on by other people. Likewise, Stephen also had 2 more chapters in his thesis which lacked proper mathematical proof’s and were loaded with full of imaginations and illusions.But, later on, he himself corrected his own theory in his next project by taking a step against his proposed theory. Gradually I started remembering an incident which occurred in my class, during my IIT-JEE practice. It was closely related to the title of my script. “Knowing a concept wrongly is more dangerous than skipping a concept”, said by my teacher, this was totally related to the negative marking of my examination. Because answering a question wrongly would fetch me negative marks. This is worse than getting zero marks for that respective question by not answering it. This small theory has made a lot of changes in my life, right from judging a person wrongly till being a person who is able to observe and learn many things from minute incidents happening in and around me.

The main point lies in the perspective, through which we see this beautiful universe. That is where the major difference lies among people, the mindset! In this world, there is a simple meaning and also a deeper meaning for all the great quotes said by most successful people around the world. One of that is, which every common man can understand and the other lies in an invisible path. One who had come across the same incident as like writer would be able to interpret the inner meaning of it.The simple version of the statement ‘’The greatest enemy of knowledge is not ignorance, it is the illusion of knowledge’’, is that “It simply states that ignoring may not harm you as much as partial or incomplete knowledge may do”.Some people think they can help others with incomplete knowledge but this would only create deeper problems than a solution.Stephen Hawking never said this quote, this was a quote from Daniel Boorstin, an American historian who was referring to how people cease to look for further information if they are arrogant enough to believe that have all they need.

Many success stories heard by us emerged from a huge misconception…with this, we can’t say that everything which is existing on this planet is purely correct. But, there is no harm in imagining a concept, perhaps the problem arises when it is publicized with all the people and it is wrong. However, every school student is advised to build the quality imagination right from their childhood for making things simpler and interesting. But, the conditions apply asterisk tells us the imagination or the illusion must be the correct path or the actual way. The one information which is being said must be accurate and correct up to the maximum extent.This quote plays an important role in our daily life also, but the thing lies at how we become aware of it and how do we respond to it!

Finally, develop a habit of updating your knowledge or facts openly. It may help you in some point of your life. Rather than being completely ignorant, this is fine. If you ignore everything, then the world will ignore you.