Descriptive Analytics Explained

What is Descriptive Analytics?

• Descriptive analytics interprets the historical data to understand changes in a business.
• It uses historical data to enable comparisons with reporting periods for the same company, quarterly or annually.
• Metrics include Year-over-year (YOY) pricing changes or month-over-month sales growth and total revenue/subscriber.
• These metrics describe what occurred in a business during a set time.

How Descriptive Analytics Functions

• It takes a full range of raw data and parses it to draw conclusions useful to managers, investors and stakeholders.
• Provides an accurate picture of past performance, compared to other comparable periods.
• It compares performance with others in the same industry.
• Performance metrics flag areas of strength and weakness to inform management strategies.
• Sales of $1 million may sound impressive but if it is a 20% month-over-month decline, there is cause for concern.
• For a 40% YOY increase, then it suggests that something is going right with sales or marketing strategy.
• An informed view of company sales performance requires larger context, including targeted growth.
• It is one of the most basic pieces of business intelligence companies use.
• While often industry-specific, is broadly accepted throughout the financial industry

Uses of Descriptive Analytics

• Companies use it to gain valuable insight into their performance.
• Companies use descriptive analytics to compare their performance and position in the marketplace with its competitors via past performance.
• Assists in determining current financial trends, like individual goals in a company.
• It can be used in different parts of any business, as it helps to understand company performance and inefficiencies.
• Corporate management identifies areas for improvement and motivates teams to implement successful changes.
• Two primary methods for collecting data are data aggregation and data mining.
• Before data is made sense of, it gathers data and parses it into manageable information.
• Management uses the information to comprehend where the business is.

Descriptive Analytics & Return on Invested Capital

• Return on Invested Capital (ROIC) is a form of descriptive analytics.
• It assesses net income, dividends, and total capital.
• The data turns data points into an easy-to-understand percentage.
• That metrics can be used to compare one company's performance to others.

Steps in Descriptive Analytics

• Identifying metrics for analysis is an important first step, determine which metrics to review: quarterly revenue or annual profit.
• Locating data necessitates identifying all internal and external sources, including databases.
• Compiling data follows identifying and locating data; formatting all data into a single, accurate format is key.
• In data analysis analyzing datasets and figures, using different tools.
• Present data utilizing visual aids like charts, graphics and videos, to provide analysts, investors and management with insight.

Advantages of Descriptive Analytics

• It disseminates information and offers major stakeholders a way to understand complex ideas, by using visuals.
• Providing stakeholders the ability to see how their company compares to others in the same industry.
• Variables such as production costs, revenue streams, and product offerings provide areas for improvement in business plans.

Disadvantages of Descriptive Analytics

• It helps to understand past performance, but it does not predict what to expect in the future.
• Market forces, changes in supply and demand, and economic swings determine the future.
• Stakeholders may find it challenging to read between the lines.
• Stakeholders choosing to analyze favorable metrics and ignore others can skew results.
• It could portray a sense that a company is more favorable than it is.

Descriptive Analytics: Pros and Cons

• Pros: Breaks down information, allows for ease of understanding, allows companies to see how they're doing against competition
• Cons: Doesn't forecast, stakeholders can handpick "favorable" metrics

Descriptive Compared to Other Analytics

• Newer analytics include predictive, prescriptive, and diagnostic analytics.
• These analytics use descriptive analytics, and integrate more data from other sources to model outcomes in the near term.
• These analytics go beyond providing info and assist in decision-making.
• They maximize positive outcomes and minimize the negative ones.

Predictive Analytics

• Predictive analytics makes predictions, via Stats and modeling, on future performance.
• Utilizes current and past data to determine whether similar outcomes are likely to happen again.
• It helps to identify and address inefficiencies.
• Also finds better and more efficient ways to put their resources (like supplies, labor, and equipment) to work.

Prescriptive Analytics

• Prescriptive analytics facilitates technology use to analyze data for specific results.
• It considers situations, resources, and past and current performance accounts, suggestions for the future.
• Better decision-making across timelines, including whether to invest more in research and development or enter a new market.

Diagnostic Analytics

• It determines variable relationships coupled with why certain trends exist.
• Helps determine why something transpired using the help of computer software or manually.
• Does not understand historical performance or make future predictions, but figure out the root cause.
• Enables changes for the future

Descriptive Analytics Use Cases

• Attempts to answer the question of "What Happened", via historial data to understand changes.
• Companies draw comparisons with reporting periods or similar companies.
• Companies identify inefficiencies and make changes for the future.

Relationship Between Descriptive & Predictive Analytics

• DA: what happened?
• PA: what will happen?
• DA will uses historical data and past performance to make improvement
• PA will try to understand those changes for future performance.
• Both work in tandem

Example Descriptive Analytics

• DA helps analyze various metrics to help a company achieve success, including measuring engagement with audiences.
• Can measure how clicks and likes lead to traffic to sites and increase sales and referrals.

Algorithms

• Clustering: Grouping similar data with K-means or hierarchical clustering, to find meaningful segments
• Association Rules: Discover relationships via Apriori and FP-Growth for market basket analysis
• Time Series: Reveal time-dependent patterns with ARIMA and exponential smoothing models
• Text Mining and NLP: Analyze customer reviews via sentiment analysis and topic modeling.
• Decision Trees: Classify and identify critical features with ID3, C4.5, and CART algorithms
• GIS: Mapping data to find patterns
• Regression Analysis: Model relationships.
• Data Mining: Anomaly, pattern recognition, identify significant patterns.

Descriptive Analytics Process

• This analysis can be divided into key steps that play a crucial role, each of which is instrumental in gleaning meaningful insight from data.
• Data Collections: The process begins by securing pertinent data from assorted origins.
• Cleaning & Preparation: This all encompassing stage entails validating data precision while addressing discrepancies such as absent values, irregularities and duplicates.
• Segmentation: Segment the data to allow a more focused analysis; helps uncover insights specific to each segment.
• KPIs: Calculates Key performance indicators for organization.
• Trend analysis: Understand how variables or metrics has changed over time.
• Data Reporting and Visualization: Provides insight into the descriptive analytics process effectively.
• Monitoring and Iteration: Requires regular data monitoring and updates, as needed.

Sales Performance Analysis with Descripitve Analytics

• Identify top selling products as wells as identifying the impact of strategies used and underperforming products
• Possible actions: adjust pricing, launch marketing, expand relevant product offerings

Customer Segmentation with Descripitve Analytics

• Target marketing campaigns, personalize communication, and customize offering

Website Analytics with Descripitve Analytics

• Help understand user behavior, such as page views, bounce rates, optimize wesbite design and improve rates.

Operational Efficiency

• Pinpoint aread and inefficiencies, by analyzin KPIs, like product times and resource usage

Financial Analysis with Descripitve Analytics

• Examine financial statements, understand trends in finances and assess helath of business.
• Can provide info, with which action, can be taken.

Qualitative and Quantitative Data Types

• Qualitative deals with data as descriptions like color, size, etc.
• Quantitative deals with numbers, such as statistics, poll numbers, percentages, etc.

Data Collection: Primary Data

• Primary data refers data directly collected from fist-hand sources.
• Can be collected via interviews, surveys, experiments, etc.
• It's current, relevant and tailored.

Direct Personal Investigation

• Is making direct contact with the person to gain info, directly
• Example: women an their daily lives

Indirect Oral Investigation

• Investigator does not make direct contact with a person.
• Data orrally collected from someone else
• Example: Superiors collecting data on employees

Advantages & Disadvantages

• Advantage: Provides natural real time data
• Disadvantage: Observer bias, limited what can be seen
• Use case: Behavioral studies

Questionnaires

• Set of questions given in person, online or on paper
• Two ways

Mailing Methods

• Investigator attaches mailed letter, explain puporse of study.

Enumerator's Method

• Prepare what's needed and reached out to informants with prepared questionairre.

Limitations

• May be biased
• Low response time

Observation Methods

• Watch behavior, events, and recoding events as they naturally occur.
• Advantages - Provides real-time, authentic data
• Disadvantage: can influence results
• Suitable Use - Studying user interactions

Experiment Methods

• manipulating on more variable and determining treatment variable compared outcomes
• Advantages: Allow cause and effect relation
• Disadvantage: Artificial
• Suitable Use - Testing efficacy of drug

Focus Group

• Gathering to discuss topic, facilitator ask open quesiton.
• Adv: In-depth insight
• Dis: Non-representative sample size
• Suitable Use- Feedback marketing

Local Sources

• Investigator appoints local persons, who are then furnished by them.

Secondary Data

• information that has already been collected, existing research papers, govt reports, books, etc.
• Its readily available and less costly. Saves time.

Government Publications

• The govt releases documents from different ministries. Examples: Stat Abstracts

Semi-Gov Publications

• Data related to health, births and deaths.

Published Sources

• Trade Associations- collect published from various sources.
• Journal and papers- different mags

International Publications

• Like IMF
• UNO
• ILO
• World Bank publish data.

Principled Differences

• Objective- specific reasons, purpose adjustments
• Originality
• Cost collection- costs more

Data vs. Intuition

• Data driven allows real time insights and predictions, compared over 'gut feelings'

Data Driven Success=

• Customer satisfaction, better planning, customer engagmenet.
• Helps retailers market campaigns and enhance suggestion engines.
• Use customer data for Dynamic pricing.

Increase Customer Retention

• Streaming service that personalizes recommendations for viewers.
• Uses viewing history, ratings, tailored recommendations. This all uses custom algorithms for recommendations.
• Helps further retain customer and reduce churn.

Proactive business practices

• Help anticipate trends.
• Finance learns fromML, to prevent fraud.
• Utility does ML to estimate energy consumption for various factors to help on deamdn forecasting.

Site Selection

• Global coffee brand opts select site based on GIS to analyze demographics.

Growth Opportunities

• Retailers analyze market dynamic and segments to to innovative products to get to segments
• Allows the refinement of strategies

Inventory Management

• Multinational retailer manage particular inventory to prevent natural disaster, by using historical sals to stock up a lot in anticipation of higher demand.

Bias Handling

• Energy company implement debasing techniques, establisihing programs to raise awareness.

Data Objective

• Steps in this process is to clearly articulate what they are.

Data Identification

• Set clear objectives
• determine needs
• evaluate sources

Organize Data

• Here structured allows new patrons.
• Cleaning data help protect data.

Performance Data Analysis

• Helps inform strat

Six Steps to Data Driven

• Implement evaluate- validate insights and measure outcomes.

Challenges of Data Driven

• Navigate quality control, ensure security and privacy, address issues
• Poor communication of insights is a challenge.

Data Analysis Types

• Descriptive aims to describe and summarize data for past performance for summaries on sales reports and surveys.
• Diagnostic- what and why events occurred for correlation.
• Predictive- forecasts trends based data.
• Prescriptive goes further.

Data Preparation: Cleaning

• Known as cleansing for improving qual
• Goal: ensure the data. -Duplicate errors. -Value errors.
• Error Syntax

Clean Data is important because....

• Integral component of work flow.
• Adopt to AI
• Al Help to streamer Data driven decisions.

Benefits of data Cleanning

• Decision making
• Enhanced performance.
• Improved Consistency

Garbage in/Garbage out

• Quality,
• Poor Data=Mistakes
• Clearer Data=Good Decisions

Data Cleaning techniques

• Standardization
• Outliners
• Duplication
• Validations

Standardization

• Help ensures conformity for accurate analysis.

Outliers

• Should be evaluated whether they were data errors or meaningful values.

Deduplication

• Streamline the process for data

Missing values

• May replace them
• Flag

Final Validation

• Verify cleanliness

Data Quality

• Its accuracy, completeness, consistency, and reliability.

Data Profiling

• The technique, help assess what's in all, including consistency.

Handling Lost Data

• Crucial aspect of Data preparation.

Types of Data Lost

• MCAR: Missing is random, not dependent on other values.
• MAR: The missigness if related to the set
• MNAr: Related to reason why data sets is missing

Imputation Method

• Replace w/MEAN/MEDIAN/Mode
• K-nn imputation
• regression

Methods - Deletion

1 Listwise- removing records that cant contian information 2 Pairwirse - available data but igonring analysis,

• algorithim approaches

Algorithmic Approaches

• EM- Estimate missing
• Random Forest = handle missing,

Advaned Tech

• Multiple inputing multiple things at once,

outlier

• Significant can influence.
• Identifying is th ecritical steo

Sta Test

• Common metrics.

Data Vs Technique

• Should not hold importance.

Ensuring Data Concistency

Consistency is important Ensure is to alignment with standards.

Data Transformation Technique

Rehning RAW, ensure

Data Scales

Is crucial is in scates.

Min Max Scaling

• Transformed into Scale

Categorical Data.

Many Machine require coding.

Advanced Cleaning techniques

• Automate- faster more efficent, data recognition
• Machie leaning more important for cleaning for data

Data Aggregation

• Process raw data source,

Data Aggregation Benefits

• Optimize data performane
• Helps specialize and quickens.

Steps

1. DATA
2. Collection PROCESSSING
3. Data Summarization

Time Agg

• Summaruzing source of data over specified period

Spatial

• Data over resources during period
• Monther conversion rates.

Manual Agg

• Manually consolidate for smaller data sets but prone to errors

Automated Agg

• Leverage a software to make collecting data easier

Data SUmariztion

• Reducing complex sets with summarization

Techniques of Smarization

• Aggregation
• Saampling
• Clustering

Benefits of technique

• Improived decision
• Time Savings
• Imporved Accuracy

Choice in Technique

• Based on customer use

Overview In summary

• One is fundamental is
• filtering all helps.

Data Visualzation

Highly effective for summarizing with charts, graphcs. Etc

Data Reduction

• High dimensional data in lower representation- Helps.

Tech Summarization

• Text Mining- text to summary

Measures

• Mean represents average, influenced by outliers
• Meridan- Is the middle value, less affected by the outliers
• Models, representing the most popular value

Range-Variance Standard

• Standard deviations.
• Data set to height

Data

• Barcharts
• Bar graphs.

Data Summaries

• Summaruzing with time setries

Data Summarization

• Crucial in extracting from business as to deeper for undert

Business Insight

• The company uses such to optimize and plan management plans

Market Research

Data is also used in market research as the company use reviews

Tech Research

Data to accelerate

Data Ethics

• Respect and Accountability.
• Data driven

Importance

• Trus,Reputation
• Compleitence and Legal
• Social