Welcome to ClearScape Analytics Experience. This service consists of multiple demonstrations of the industry leading in-database analytics that you can run on your own. You can modify them or use them as examples to use with your own tools against our data or small (not sensitive) data you upload. Each notebook will:
- describe the business situation,
- will attach the needed data from the cloud, and
- walk you step-by-step through the use of the ClearScape Analytics functionality.
These are functional demonstrations executed on a tiny platform with small data, but the same functionality is available on all of our platforms up to one with hundreds of nodes and petabytes of data. ClearScape Analytics allows you to apply AI, ML and advanced statistics to your data without the cost and complexity of exporting data. You can develop sophisticated models on other platforms with your favorite tools and import those models to execute in production at massive scale.
If you've never used Jupyter before, we strongly recommend reviewing the First Time User section of Getting Started. You'll find an introduction video with tips on using this platform. There are also tips for you if you just want to look without programming. If you have questions or issues, click here to send an e-mail to ClearScape Analytics Support.
Items in italics are coming soon.
| Getting Started | Industries | Business Function | Analytic Function | 3rd Party Tools |
|---|---|---|---|---|
| First Time User | Automotive | Finance | Data Preparation | AWS SageMaker |
| I am not a programmer | Energy & Natural Resources | Marketing | Descriptive Statistics | Azure ML |
| Developer Information | Financial | Feature Engineering | Dataiku | |
| Healthcare | Generative AI | H2O.ai | ||
| Manufacturing | Geospatial | Microsoft PowerBI | ||
| Retail | Hypothesis testing | MicroStrategy | ||
| Telco | Machine learning | R | ||
| Travel & Transportation | ModelOps | SAP Business Objects | ||
| Object Storage | SAS | |||
| Open-and-connected analytics | Tableau | |||
| Path Analytics | Vertex | |||
| Text Analysis |
Video description how to find demos in the index and folder view, tips on running demos and options for foreign vs local tables used in the demonstrations in your ClearScape Analytics environment.
Information Only
When running a Jupyter Notebook, there are various indicators that show what is happening. This is a guide to those indicators.
SQL Version Read Only SQL Version
Not everyone that uses this site will want to learn programming. Some will want to review the business cases, look at the steps for the analysis and look at the tables, charts and maps. This is a guide for those people.
Information Only
Shows how to use python to load CSV data from local storage and from zipped files
Python Version
Shows multiple ways to load data from local CSV files, and cloud files on Google and AWS using the SQL kernel.
SQL Version
Demonstration of using REST API calls to Vantage which is useful for web or mobile applications to access and maintain data.
Python Version
This guide will walk you through writing your first SQL queries in Jupyter. It uses some of the Vantage system tables as a source for the queries.
SQL Version Read Only SQL Version
Introduction to Teradataml package for Python including connecting to Vantage, Teradata DataFrames, data manipulation and export to Pandas.
Python Version Read Only Python Version
Provides step-by-step instructions on the basics of using Python Pandas with Jupyter notebooks.
Python Version Read Only Python Version
Data from queries is brought to life with graphics and charts. This shows how to use the %chart magic command to display results.
SQL Version Read Only SQL Version
Vantage Analytics Library (VAL) is a set of over 50 functions for advanced analytics. This provides an overview and links to an 8 minute video overview.
SQL Version Read Only SQL Version
This provides an index to all of the databases used by demo notebooks on this machine, allowing you to use that data for your own notebooks or BI tools.
Python Version
It is very easy to submit your demo for publication. Tell us directory with the notebook and referenced files and grant us access to your database. We'll take it from there.
Python Version
Evaluates potential failures in spot welds based on voltage anomalies during the welding process.
Python-SQL Version Read Only Python-SQL Version
Uses local data (or foreign tables on GCP) to analyze patterns of battery failure, then links to data on AWS for detailed battery measurement.
Python-SQL Version Read Only Python-SQL Version
This combines machine learning and BYOM to forecast energy consumption using Vantage to score the model at scale without having to export all data.
Python Version
Demonstration of using Dataiku with Vantage. Instructions provided for use with your Dataiku copy + screen shots if you don't have Dataiku. PMML model from Dataiku is imported to Vantage for execution and scoring.
Python Version Read Only Python Version
Uses a combination of Vantage Analytics Library to prepare data, using machine learning in python and importing the resulting PMML model into Vantage for scoring.
Python Version Read Only Python Version
Predicts the future demand for cast in Automatic Teller Machines (ATMs) using Auto Regressive Integrated Moving Average using In-Database functions.
SQL Version
Uses government consumer complaint data with SQL queries and visualizations to identify source of top complaints.
SQL Version Read Only SQL Version
This shows the use of ClearScape analytics to reduce the pre-processing effort of incoming raw credit card data to prepare for analysis of potential loan defaulters.
SQL Version Read Only SQL Version
Uses analytic techniques to find new customers, measure marketing attribution, and maximizing marketing effectiveness
SQL Version Python-SQL Version
Detect financial transaction fraud using powerful in-database machine learning functions
Python Version Python-SQL Version Read Only Python Version Read Only Python-SQL Version
A model that was developed externally is imported into Vantage for evaluation and execution at scale to detect fraud.
Python Version Python-SQL Version Read Only Python Version Read Only Python-SQL Version
Show As-IS/As-Was capabilities of Vantage Temporal to dramatically simplify the SQL and improve performance for analyzing insurance policies versus claims.
SQL Version Read Only SQL Version
This is a guide through the PIMA Diabetes prediction including data exploration and model experimentation.
Python Version
Covers the ModelOps operationalizing of Bring-your-own-model (BOYM) importing a model via PMML. PMML allows exchange predictive models produced by data mining and machine learning algorithms.
Python Version
Covers the ModelOps operationalizing of the ONNX model format for BYOM for the Diabetes use case. ONNX is an efficient model format primarily related to neural networks.
Python Version
Covers the ModelOps operationalizing of the H2O model format for BYOM for the Diabetes use case. H2O is an open source, distributed in-memory machine learning library with linear scalability.
Python Version
This uses BYOM to import a trained H2O model to identify potential diabetes patients. BYOM allows the data scientist to create models in languages they prefer and run at scale inside Vantage without moving data.
Python-SQL Version
Decision Forest and Generalized Linear Model are applied to identify factors that indicate diabetes. The earlier the disease is identified, the better the chance of reducing organ damage.
Python Version Python-SQL Version
Machine learning is applied to the complex attributes of patients to help recognize patterns that may lead to heart failure faster than a human may recognize.
Python-SQL Version Read Only Python-SQL Version
The ClearScape Analytics Attribution function is used to determine the weight of various events that precede the final outcome, in this example, knee replacement.
SQL Version Read Only SQL Version
This uses the ClearScape Analytics nPath® function to provide visuals on the events leading up to the final outcome, in this case, knee replacement.
Python Version Read Only Python Version
Evaluates potential failures in spot welds based on voltage anomalies during the welding process.
Python-SQL Version Read Only Python-SQL Version
Uses local data (or foreign tables on GCP) to analyze patterns of battery failure, then links to data on AWS for detailed battery measurement.
Python-SQL Version Read Only Python-SQL Version
Uses geospatial techniques to locate service centers close to the complaint and outlier detection to detect part defects earlier than expected.
Python-SQL Version Read Only Python-SQL Version
Uses analytic and ML techniques to predict how long vehicle testing will take based on combination of features installed.
Python-SQL Version Read Only Python-SQL Version
Uses ML functions to predict failures to identify and mitigate potential machine failures before they occur.
Python-SQL Version Read Only Python-SQL Version
Applies machine learning to predict Remaining Useful Life (RUL) of jet engines, allowing scheduling of maintenance and replacement before failure occurs and reduces the cost of maintenance and replacement.
Python-SQL Version Read Only Python-SQL Version
This uses nPath® analysis to show the retail events that prevent the ultimate objective: a purchase. This uses interactive Sankey diagrams to understand the problems.
Video Version Python Version Read Only Python Version
Analysis of customer purchase behavior using nPath® analysis in Python with visualization using Sankey diagrams.
Python Version Python-SQL Version Read Only Python Version Read Only Python-SQL Version
This demonstrates integration of local data or foreign tables on GCP and integration across cloud providers to detailed historical sales records on AWS.
SQL Version
Hyper-personalization creates models from customer interations on multiple channels to determine the "Next Best Offer" for the individual.
Python-SQL Version Read Only Python-SQL Version
This uses the unsupervised K-Means ML algorithm to identify patterns in retail purchases.
SQL Version Read Only SQL Version
This is a somewhat simplified version of the Multi-Touch Attribution demonstration focused on the interests of the Business Analyst vs the Data Scientist.
Python-SQL Version Read Only Python-SQL Version
Demonstrates attribution of customer behavior via single touch and multi-touch rule-based models, and using statistical, and algorithmic models. Multiple approaches are demonstrated since each has strengths and limitations.
Python-SQL Version Read Only Python-SQL Version
This creates an ARIMA time series model based on holidays and merchandising activities impacting store sales for a hypermarket retailer.
Python Version Read Only Python Version
Predicts demand for retail products showing how multiple models can be run concurrently.
Python-SQL Version Read Only Python-SQL Version
Forecasts total store sales using ARIMA (AutoRegressive Integrated Moving Average)
Python Version Read Only Python Version
Uses the Script Table Operator (STO) to run the H20 machine learning library as an extension to ClearScape Analytics. H20 runs in parallel inside the Vantage database without exporting data to another platform.
Python-SQL Version Read Only Python-SQL Version
Use NGram splitter to analyze comments retail products to determine patterns of words used to describe products.
SQL Version Python Version Python-SQL Version
Uses graph analysis to identify communities and key influencers within Call Data Records. This uses Script Table Operator to invoke external procedures to work inside the Vantage database without exporting data.
Python-SQL Version Read Only Python-SQL Version
This uses logistic regression for supervised learning to predict the probability of a customer switching vendors based on usage patterns, billing information demographics and interactions. XGBoost is then used in database to improve the prediction.
Python-SQL Version Read Only Python-SQL Version
Demonstrates the ability of Geospatial to show signal strength, coverage areas and travel path of customers through cell tower coverage area.
Python-SQL Version Read Only Python-SQL Version
Demonstration of Geospatial and TimeSeries using Austin bike trip data between 2014 and 2017.
SQL Version Python-SQL Version
Applies Auto Regressive Integrated Moving Average (ARIMA) analysis to forecast airplane passenger volume.
SQL Version Read Only SQL Version
Applies geospatial functions to analysis of NYC Taxi pickups and drop off locations.
Python-SQL Version Read Only Python-SQL Version
Applies machine learning to predict Remaining Useful Life (RUL) of jet engines, allowing scheduling of maintenance and replacement before failure occurs and reduces the cost of maintenance and replacement.
Python-SQL Version Read Only Python-SQL Version
Uses nPath displays to show relationship of delays and predictive models to anticipate potential delays and enable proactive planning.
Python Version Read Only Python Version
Hyper-personalization creates models from customer interations on multiple channels to determine the "Next Best Offer" for the individual.
Python-SQL Version Read Only Python-SQL Version
This is a somewhat simplified version of the Multi-Touch Attribution demonstration focused on the interests of the Business Analyst vs the Data Scientist.
Python-SQL Version Read Only Python-SQL Version
Demonstrates attribution of customer behavior via single touch and multi-touch rule-based models, and using statistical, and algorithmic models. Multiple approaches are demonstrated since each has strengths and limitations.
Python-SQL Version Read Only Python-SQL Version
This shows the use of ClearScape analytics to reduce the pre-processing effort of incoming raw credit card data to prepare for analysis of potential loan defaulters.
SQL Version Read Only SQL Version
This demonstrates a subset of the over 100 analytic functions in the teradataml package for Python
Python Version Python-SQL Version
Outliers in an analysis skew the results and make it difficult to recognize the main patterns. ClearScape Analytics has tools to remove outliers automatically.
SQL Version Read Only SQL Version
This performs in-database analysis of data values, distribution, histograms, and text field analysis using SQL to access the Vantage Analytics Library.
SQL Version Read Only SQL Version
Demonstrated the use of Teradataml in Python to perform descriptive statistics, transformation, model building model evaluation and scoring.
Python Version Read Only Python Version
Demonstration of Geospatial and TimeSeries using Austin bike trip data between 2014 and 2017.
SQL Version Python-SQL Version
Uses geospatial techniques to locate service centers close to the complaint and outlier detection to detect part defects earlier than expected.
Python-SQL Version Read Only Python-SQL Version
Applies geospatial functions to analysis of NYC Taxi pickups and drop off locations.
Python-SQL Version Read Only Python-SQL Version
Creative application of geospatial to locations of sensors in a research lab and integration of data from tables with detailed recordings on cloud storage.
Python-SQL Version Read Only Python-SQL Version
Demonstrates the ability of Geospatial to show signal strength, coverage areas and travel path of customers through cell tower coverage area.
Python-SQL Version Read Only Python-SQL Version
This demonstrates a subset of the 18 hypothesis test in the Vantage Analytics library using SQL such as Parametric, Binomial, Kolmogorov/Smirnoff, Rank, etc.
SQL Version Read Only SQL Version
Covers the ModelOps operationalizing of Bring-your-own-model (BOYM) importing a model via PMML. PMML allows exchange predictive models produced by data mining and machine learning algorithms.
Python Version
Covers the ModelOps operationalizing of the ONNX model format for BYOM for the Diabetes use case. ONNX is an efficient model format primarily related to neural networks.
Python Version
Covers the ModelOps operationalizing of the H2O model format for BYOM for the Diabetes use case. H2O is an open source, distributed in-memory machine learning library with linear scalability.
Python Version
Evaluates potential failures in spot welds based on voltage anomalies during the welding process.
Python-SQL Version Read Only Python-SQL Version
Uses a combination of Vantage Analytics Library to prepare data, using machine learning in python and importing the resulting PMML model into Vantage for scoring.
Python Version Read Only Python Version
Decision Forest and Generalized Linear Model are applied to identify factors that indicate diabetes. The earlier the disease is identified, the better the chance of reducing organ damage.
Python Version Python-SQL Version
Detect financial transaction fraud using powerful in-database machine learning functions
Python Version Python-SQL Version Read Only Python Version Read Only Python-SQL Version
A model that was developed externally is imported into Vantage for evaluation and execution at scale to detect fraud.
Python Version Python-SQL Version Read Only Python Version Read Only Python-SQL Version
Uses graph analysis to identify communities and key influencers within Call Data Records. This uses Script Table Operator to invoke external procedures to work inside the Vantage database without exporting data.
Python-SQL Version Read Only Python-SQL Version
Uses analytic and ML techniques to predict how long vehicle testing will take based on combination of features installed.
Python-SQL Version Read Only Python-SQL Version
Machine learning is applied to the complex attributes of patients to help recognize patterns that may lead to heart failure faster than a human may recognize.
Python-SQL Version Read Only Python-SQL Version
Hyper-personalization creates models from customer interations on multiple channels to determine the "Next Best Offer" for the individual.
Python-SQL Version Read Only Python-SQL Version
This uses the unsupervised K-Means ML algorithm to identify patterns in retail purchases.
SQL Version Read Only SQL Version
This is a somewhat simplified version of the Multi-Touch Attribution demonstration focused on the interests of the Business Analyst vs the Data Scientist.
Python-SQL Version Read Only Python-SQL Version
Demonstrates attribution of customer behavior via single touch and multi-touch rule-based models, and using statistical, and algorithmic models. Multiple approaches are demonstrated since each has strengths and limitations.
Python-SQL Version Read Only Python-SQL Version
Uses ML functions to predict failures to identify and mitigate potential machine failures before they occur.
Python-SQL Version Read Only Python-SQL Version
Applies machine learning to predict Remaining Useful Life (RUL) of jet engines, allowing scheduling of maintenance and replacement before failure occurs and reduces the cost of maintenance and replacement.
Python-SQL Version Read Only Python-SQL Version
Predicts demand for retail products showing how multiple models can be run concurrently.
Python-SQL Version Read Only Python-SQL Version
Forecasts total store sales using ARIMA (AutoRegressive Integrated Moving Average)
Python Version Read Only Python Version
Uses the Script Table Operator (STO) to run the H20 machine learning library as an extension to ClearScape Analytics. H20 runs in parallel inside the Vantage database without exporting data to another platform.
Python-SQL Version Read Only Python-SQL Version
This uses logistic regression for supervised learning to predict the probability of a customer switching vendors based on usage patterns, billing information demographics and interactions. XGBoost is then used in database to improve the prediction.
Python-SQL Version Read Only Python-SQL Version
Uses nPath displays to show relationship of delays and predictive models to anticipate potential delays and enable proactive planning.
Python Version Read Only Python Version
Demonstration of Vantage Analytic Library scoring and analytic functions like linear regression, decision trees, K-Means clustering, Factor Analysis, etc.
SQL Version Read Only SQL Version
In this tutorial, we will show how to integrate Vantage Analytics capabilities in Vertex AI ML Pipelines. Vertix AI is the environment for data scientists to deploy ML models.
Python Version
This introduction and table of contents introduces you to ModelOps and provides a launch for ModelOps. It is recommended to go through ModelOps deomstrations in sequence.
Information Only
This introduces the ModelOps methodology, provides an overview video, and a description of navigating the projects, models, and datasets plus a description of monitoring capabilities.
Python Version
This is a guide through the PIMA Diabetes prediction including data exploration and model experimentation.
Python Version
Covers the ModelOps operationalizing of Bring-your-own-model (BOYM) importing a model via PMML. PMML allows exchange predictive models produced by data mining and machine learning algorithms.
Python Version
Covers the ModelOps operationalizing of the ONNX model format for BYOM for the Diabetes use case. ONNX is an efficient model format primarily related to neural networks.
Python Version
Covers the ModelOps operationalizing of the H2O model format for BYOM for the Diabetes use case. H2O is an open source, distributed in-memory machine learning library with linear scalability.
Python Version
Shows you how to set up your own GIT repository for models and create a new project in ModelOps associated with your new repository. This step is required for the next notebooks.
Python Version
For the project you've created in ModelOps, this shows definition of the training function, evaluate function, scoring function, life cycle, and monitoring.
Python Version
Demonstrates the use of ModelOps to finalize the H2O AI model, train, evaluate, approve, deploy, score and monitor.
Python Version
Uses XGBoost algorithm to generate both Python Joblib and PMML model formats and operationalize them through ModelOps.
Python Version
Uses the gbm R package to generate both R model formats and operationalize through ModelOps. The gbm R packages extends Freund & Schapire's AadaBoost algorithm and Friedman's Gradient Boosting Machine (gbm).
Python Version
Uses local data (or foreign tables on GCP) to analyze patterns of battery failure, then links to data on AWS for detailed battery measurement.
Python-SQL Version Read Only Python-SQL Version
This demonstrates integration of local data or foreign tables on GCP and integration across cloud providers to detailed historical sales records on AWS.
SQL Version
Evaluates potential failures in spot welds based on voltage anomalies during the welding process.
Python-SQL Version Read Only Python-SQL Version
Uses a combination of Vantage Analytics Library to prepare data, using machine learning in python and importing the resulting PMML model into Vantage for scoring.
Python Version Read Only Python Version
Discusses how the 3rd party tool DataIku can be used with Vantage.
Information Only
This uses BYOM to import a trained H2O model to identify potential diabetes patients. BYOM allows the data scientist to create models in languages they prefer and run at scale inside Vantage without moving data.
Python-SQL Version
This combines machine learning and BYOM to forecast energy consumption using Vantage to score the model at scale without having to export all data.
Python Version
Demonstration of using Dataiku with Vantage. Instructions provided for use with your Dataiku copy + screen shots if you don't have Dataiku. PMML model from Dataiku is imported to Vantage for execution and scoring.
Python Version Read Only Python Version
A model that was developed externally is imported into Vantage for evaluation and execution at scale to detect fraud.
Python Version Python-SQL Version Read Only Python Version Read Only Python-SQL Version
Introduction to Teradataml package for Python including connecting to Vantage, Teradata DataFrames, data manipulation and export to Pandas.
Python Version Read Only Python Version
In this tutorial, we will show how to integrate Vantage Analytics capabilities in Vertex AI ML Pipelines. Vertix AI is the environment for data scientists to deploy ML models.
Python Version
This uses nPath® analysis to show the retail events that prevent the ultimate objective: a purchase. This uses interactive Sankey diagrams to understand the problems.
Video Version Python Version Read Only Python Version
Analysis of customer purchase behavior using nPath® analysis in Python with visualization using Sankey diagrams.
Python Version Python-SQL Version Read Only Python Version Read Only Python-SQL Version
Uses analytic techniques to find new customers, measure marketing attribution, and maximizing marketing effectiveness
SQL Version Python-SQL Version
The ClearScape Analytics Attribution function is used to determine the weight of various events that precede the final outcome, in this example, knee replacement.
SQL Version Read Only SQL Version
This uses the ClearScape Analytics nPath® function to provide visuals on the events leading up to the final outcome, in this case, knee replacement.
Python Version Read Only Python Version
Shows rule-based, Statistics, and Algorithmic attribution of the marketing touchpoints leading to conversion. Ten approaches will be demonstrated along with path analysis of effectiveness and cost of conversion.
Python-SQL Version Read Only Python-SQL Version
Uses nPath displays to show relationship of delays and predictive models to anticipate potential delays and enable proactive planning.
Python Version Read Only Python Version
Use NGram splitter to analyze comments retail products to determine patterns of words used to describe products.
SQL Version Python Version Python-SQL Version
This performs in-database analysis of data values, distribution, histograms, and text field analysis using SQL to access the Vantage Analytics Library.
SQL Version Read Only SQL Version
Demonstration of Geospatial and TimeSeries using Austin bike trip data between 2014 and 2017.
SQL Version Python-SQL Version
Applies Auto Regressive Integrated Moving Average (ARIMA) analysis to forecast airplane passenger volume.
SQL Version Read Only SQL Version
A key component of ESG is Carbon Footprint. This demonstrates a part of a solution available from Teradata to integrate multiple data sources to calculate carbon footprint of various corporate activities.
Python-SQL Version Read Only Python-SQL Version
Predicts the future demand for cast in Automatic Teller Machines (ATMs) using Auto Regressive Integrated Moving Average using In-Database functions.
SQL Version
Uses government consumer complaint data with SQL queries and visualizations to identify source of top complaints.
SQL Version Read Only SQL Version
This combines machine learning and BYOM to forecast energy consumption using Vantage to score the model at scale without having to export all data.
Python Version
Fourier transformations are demonstrated to filter out noise from signals to allow identification of underlying patterns.
SQL Version Read Only SQL Version
Show As-IS/As-Was capabilities of Vantage Temporal to dramatically simplify the SQL and improve performance for analyzing insurance policies versus claims.
SQL Version Read Only SQL Version
Time series data can answer the questions about what was happening at a point in time. This applies Time series and temporal capabilities of vantage to NYC taxi data.
SQL Version Read Only SQL Version
This creates an ARIMA time series model based on holidays and merchandising activities impacting store sales for a hypermarket retailer.
Python Version Read Only Python Version
Analysis of sessions and queries you executed using the built-in logging facilities of Vantage.
SQL Version Read Only SQL Version
Discusses how the 3rd party tool DataIku can be used with Vantage.
Information Only
Demonstration of using Dataiku with Vantage. Instructions provided for use with your Dataiku copy + screen shots if you don't have Dataiku. PMML model from Dataiku is imported to Vantage for execution and scoring.
Python Version Read Only Python Version
Uses the gbm R package to generate both R model formats and operationalize through ModelOps. The gbm R packages extends Freund & Schapire's AadaBoost algorithm and Friedman's Gradient Boosting Machine (gbm).
Python Version
©2023 Teradata. All Rights Reserved