Realtime Marketing & Sales Analytics Solutions
Lambda architecture: A data processing architecture that designed to handle massive quantities of data by taking advantages of both batch and real time streaming processing methods.
Apache SPARK is the real time streaming engine that allows you to seamlessly integrate batch HDFS infrastructure and real time stream processing.
SPARK is fast, distributed, and scalable and fault tolerant system processing streaming data using SCALA programming.
Predictive analytics and NLP analysis – Twitter Firehose.
Enables low latency, high performance and handles massive load. Automatic recovery from Spark engine failures.
Similar Real time Streaming projects:
• EMR with AWS Redshift.
• Apache Kafka with Twitter Strom.
• Yahoo! S4.
• LinkedIn Samsa.
• Google Millwheel.
• Microsoft Dryad.
Real time streaming analytics: Handle all the 3Vs of Big Data in one platform.
AWS Streaming Solutions :
• Continuous analytics results from the event stream.
• Analytics with Spark over Casandra.
• Data analyzed in motion – as it arrives.
• Business functions - Monitoring, Counting, Alerting, Reporting.
• Real time Decision Making with Predictive analytics.
• Finding missed opportunities.
• Finding new markets and opportunities.
• Finding more revenues and cost savings.
• Cutting preventable losses.
Machine Learning With Live Streaming:
• Brand and product reputation on Social media – Twitter and FB.
• Ecommerce – Listening & learning from customers and Sensitive Inventory.
• Medical – Complex analytics in ICU.
• Finance – Fraudulent alert or Risky trades.
• Manufacturing – Preventive maintenance and finding process defects.
• IT – Network optimization and auto capacity management.
Are you ready for Realtime Analytics?
we’re here to help
Real-Time Analytics: Dynamic, instantaneous analysis and reporting based on the continuous input of large volumes of data. It is an instant marketing, sales and customer service decision making tool. Part of the process of delivering information about business operations as they occur.
- To provide intraday reviews.
- Make real time business decisions.
- Expedite orders are on hold that need to be shipped.
- To view real time sales and customer service data.
- Price and loyalty analytics.
- View supply chain and order management data.
- Complex calculations
- Large data volumes
Reasons for use of Real-Time Analytics:
Real-Time Analytics challenges
Elastic search is a powerful search and analytics engine. It uses document store where filed is indexed and searchable.
Documents are represented using JSON. Distributed and scalable. Data can be in nested JSON objects.
Elastic search is a Java instance and each instance referred as a node.
It exposes a RESTful API. Send and receive via JSON payloads.
Clients – Ruby, Python, .Net, Angular JS and React.
Full text SQL search Vs Elastic Search.
• SQL is a relational database. Elastic is a search engine.
• SQL filters at binary level, Elastic search thrives at binary and full text search.
• SQL indexes are built on data store, ES uses NO SQL indexing.
Field of study that gives computers the ability to learn without being explicitly programmed.
Three niches for machine learning
1) Data mining: using historical data to improve decisions: Medical records - medical knowledge.
2) software applications that are difficult to program by hand - Autonomous driving, Image classification.
3) User modeling - Automatic recommender systems.
Open Source Machine learning - Apache Mahout, Spark Mlib, H2O prediction engine.
•Your team is constantly generating content.
•You are tasked with making this knowledge base searchable and accessible.
•You need key search features including text matching, faceting, filtering, fuzzy search, auto complete, and highlighting.
Purpose of Elastic search
•Share content information.
•Extract product and customer insight.
•Recognize customer patterns.
•Track operational performance.
Ultimately, make better business, technical, and operational decisions.
Combined with Logstash and Kibana, the ELK stack provides a tool for real-time analytics and data visualization.
Routing can be used to control which shards (and therefore which nodes) receive requests to search for a document. When routing is enabled the user can specify a value at either index time or query time to determine which shards are used for indexing/querying.
The same routing value is always routed to the same shard for a given index.
Google Cloud machine learning
Customizable and fully managed service - TensorFlow graph.
Google Translate and Google Speech API.
Google Vision API -Faces, landmarks, logos, etc.
Wix Company -More accessible platform to broad aspect of people.