Waching daily May 30 2018

ZEPTO is an AI-powered Cloud based Web Application for Data Analytics

Each user will be provided a Username and a Password to log in

This section is called Data Locker

Imagine this area as your Google Drive

You can upload your Structured Excel or CSV files here

There are two sections

The left side is to manage your datasets in folder like boxes

The right side is where your uploaded datasets will get listed

Let's create a Data Box

By clicking on the name, you can simply rename the Box

You can customize the Box for easy identification as well

Now, let's upload a Dataset to get started

As soon as the data gets uploaded, user will get a popup to get Insights

For the demo purpose, let's just ignore it and continue

As soon as the data gets uploaded, user will get a popup to get Insights

Before we start, let's just preview the dataset

This data is about a sample retail shop sales data

and has about 15 columns and about 3000 rows

This data is about a sample retail shop sales data

The right section is where users can update their data

For example, if you uploaded the January data and now you want to append the February data to it,

that let you can upload here and as long as both the files are having same column headers

you can upload here and as long as both the files are having same Column Headers,

that let you can upload here and as long as both the files are having same column headers

all the records will be appended as a single dataset

By clicking on the Dataset name, you can rename the dataset

With simple drag and drop, you can organize and manage the Data Locker

Mark the data as the favorite for easy retrieval

You can simply share the dataset with another user to analyze collaboratively as well

Now, let's see some insights from the data

You can get some quick stories on your data as well as customized one

Let me show some quick insights first

Click here to connect the Dataset from Data Locker

Choose the data and click Connect

Here, you can see how many Insights have been generated

The first type of Insights show you if there are any categories dominating in a particular segment

In this case, the Majority of the Discounts are given to Regular Customers for Vegetable Sales

If you scroll over, you can see the values in the Pie Chart as well

To train Zepto's AI Model, you can mark how useful this insight is

If you already know this piece of information, you can mark as Thumbs Down,

so the System will learn and not show such insights next time

Likewise, if any insight is useful, you can mark as Thumbs Up,

so it will generate more relevant insights to give you an Analytical Perspective on the given insight

On this instance, in Mumbai,

Food and Vegetable items are sold more, while Regular Customers contribute for a majority of Discount

This gives an idea that Regular Customers prefer to buy more Food and Vegetable items

with a large amount of Discounts in Mumbai, which is a useful insight to take action upon it

The second type of insight shows Correlation between Measures

In this instance, Zepto identifies a strong Correlation between Sold Quantity and Sales Value,

which is an obvious fact, hence, you can train the model to not to show these kinds of insights ever again

Third, Zepto will identify Category Anomalies in the data

Zepto will pickup, unusually behaving categories and flag it to the users

as described earlier,

users can dig deep, based on its importance to the business

The fourth type of insight is that Zepto will statistically identify, which variables are affecting the measures most

In this case, the Price is affected by both Product and Financial Year,

which indicates a variety of Products with different Price ranges and all are affected by Inflation

Finally, Zepto will identify Trend Patterns over time,

and shows the highly trending up or down segments in your data

In this case, Discounts given on Saturdays are trending upward

If you are wondering why, you can always dig deep to get more answers

If you want to get insights on a specific scope of your data, you can use Custom Insights

Once the data is connected, all the Field Names in the Dataset will get listed

Numerics will be colored as Blue, Date as Black, and Segments or Text Fields are in Green

This is a 2 step Process

First select the Measure or Scope from the Dataset

Then, apply Filters to narrow down the Scope

In this case, we are trying to get insights on Sales in New York, for Vegetables and Personal Care segments

As you can clearly see, Sales of Vegetables and Personal Care segment

in New York is picking up on Thursdays and Regular Customers are contributing more to it

Further, it reveals that Personal Care segment is doing well in Saturdays and Fridays relatively

Now, let's move to Play Area to visually analyze the data

Click here to connect the Dataset from Data Locker

Once connected, all the Field Names will be listed here as listed previously in the Custom Insights section

The first cage is the Y axis, where you can drag and drop the Measure Fields like Sales, Quantity etc., in it

The second cage is the X axis, where you can drag and drop

the Category and Date Fields like, Location, Product, Transaction date, in it

If you wish to visualize different series in your data, you can use the third cage

All the Filters which are applied to the Chart will be listed in this area

There are 12 different most commonly used visualization types are provided in the Product

This would mostly cover the majority of the analysis performed on a day-to-day basis

Now let's create a Bar Chart,

showing Total Sales,

by Transaction Date,

aggregated by Year

To change the date aggregation,

click this drop down and select the desired level

Users can easily switch between the Chart Types to check the same analysis in a different visualization

By checking this tick box, users can get a Time-Series Forecasting on their data

Traditionally, the Forecasting Models are linear trendlines;

however, this Predictive Model is able to detect the Seasonal Pattern in the data and mimic the same in the forecasting as well

With few clicks, you can filter in or out across different categories to visually analyze your data

You can sort the Chart by either Labels or Values of the Bars;

in this way, you can focus on the categories which require the most attention in a snap

Once the chart is created, you can rename it and save in the Chart Vault

Chart Vault is where all the Analysis will be stored for future use

Once the Chart is saved here, users can easily share it with other users to collaboratively analyze

Once your analysis is done, let's say you want to monitor this over time

With few clicks, you can push this analysis to the Dashboard

and create your own Dashboard to monitor your business

This is a sample Dashboard

Each user can create up to 5 Dashboards and each Dashboard can accommodate 6 Chart Tiles in it

In the Tile Settings, users can set, how the Chart should change with the streaming or updating data

Each Dashboard can be shared with other users,

so all the team members can have an updated view on the business