Enterprise Scale Disease Modeling Web Portal
i Last Updated: 5/8/2015 4:13 PM3/5/2015 10:00 AM
PROPOSAL
To Develop an
Enterprise Scale Disease Modeling Web Portal
For Ascel Bio
Updated March 2015
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Guidelines and tips
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
web portal proposal in developing a web based system, Study Guides, Projects, Research of Information Systems
its to guide on how to come up with proposal
Typology: Study Guides, Projects, Research
2019/2020
1 / 9
This page cannot be seen from the preview
Don't miss anything!
Related documents
Partial preview of the text
Download web portal proposal in developing a web based system and more Study Guides, Projects, Research Information Systems in PDF only on Docsity!
Enterprise Scale Disease Modeling Web Portal
i Last Updated: 5/8/2015 4:13 PM3/5/2015 10:00 AM
PROPOSAL
To Develop an
Enterprise Scale Disease Modeling Web Portal
For Ascel Bio
Updated March 2015
Enterprise Scale Disease Modeling Web Portal
ii Last Updated: 5/8/2015 4:13 PM3/5/2015 10:00 AM
- INTRODUCTION TABLE OF CONTENTS
- 1.1 PROJECT OVERVIEW
- 1.2 BACKGROUND
- 1.3 OBJECTIVE
- PROPOSED TECHNICAL APPROACH
- 2.1 PHASE 1 REQUIREMENTS
- 2.1.1 DELIVERABLES
- 2.2 PHASE 2 REQUIREMENTS
- 2.2.1 DELIVERABLES
- 2.3 ARCHITECTURE DESIGN
- 2.4 IMPLEMENTATION
- 2.5 PROJECT CHALLENGES AND RISK MANAGEMENT
Exigence analyses text and estimates the probability that a disease is causing significant social disruption. Fatalities, people fleeing the outbreak area, and isolation wards are some of the indicators that Exigence brings to the attention of the analyst. Exigence can also be used to generate a file that Delphi can use to forecast a disease outbreak. A typical Exigence modeling workflow:
- User launches Exigence on their desktop computer.
- Exigence periodically extracts text from articles posted on select RSS Feeds.
- User runs model.
- Exigence generates a Microsoft Excel file containing a list of social disruption indicators and their probabilities.
1.3 OBJECTIVE
Ascel Bio would like to move to a software as a service (SaaS) delivery model. SaaS would bring a number of benefits besides eliminating the need for customers to install software. Computations would be done on Ascel Bio servers, allowing more complex models to be developed and executed against larger datasets than is currently possible on the customer’s computer. Delphi and Exigence were built using VBA, which is based on a software development paradigm incompatible with SaaS. Therefore, only the algorithms from the existing Delphi and Exigence applications can be reused in the new SaaS system.
The new SaaS web portal will allow customers to quickly purchase an account that gives them access to the epidemic forecast statistics corresponding to their location of interest. These locations can be their local city or other cities that they subscribed too.
A completely new Ascel Bio website will be developed to support the marketing of the new modeling portal. The main public landing page will provide a selected set of disease forecasts that are based on non-proprietary health data. Disease prevention tips and news articles will also be available on the website. The main portal will display the disease forecasts for only a few of the diseases. These selected diseases will be changed periodically based on Ascel Bio’s interest. A subscription page, accessed from the main public landing page, will provide disease forecasts for a wider variety of diseases, at the subscriber’s request, for specific targeted locations.
2. PROPOSED TECHNICAL APPROACH
Ascel Bio is a small, but growing startup. The data analysis system must therefore allow the on- demand scalability of computational resources and data storage capacity in both the short and long term. The system must be able to simultaneously run a large number of different types of models, such as machine learning and standard statistical models. In addition, adding, removing, and modifying the existing models should be supported.
We propose to meet the above requirements by developing a data analysis system based on tools already available in the Big Data ecosystem. The Big Data paradigm allows for horizontal scaling of both data storage and computational capacity using commodity machines. The paradigm is inherently modular, allowing models to be quickly added and removed from a deployed system. The costs associated with software licensing will also be minimal as most Big Data software are open source and allow free commercial usage. The Big Data stack will be responsible for data ingestion, storage, and model execution. A website development stack will also be utilized to develop the web portal. Similar to Big Data software, there is an abundance of free, open source licensed webservers and web frameworks.
As requested, system development will be comprised of two separate phases: Phase 1 and Phase 2. Each of these phases will add more functionality to the pre-existing system. A phase will consist of the four major steps: design, development, test, and production deployment. Upon the acceptance of this proposal, a detailed project plan will be created that specifies the tasks within each of those steps.
System maintenance and support are not covered under this proposal. Nor are the costs of hosting the system. If invited to, a subsequent proposal covering system hosting, maintenance and support will be created.
2.1 PHASE 1 REQUIREMENTS
The initial design and build out of the hardware, and software infrastructure used in all subsequent phases will be the main objective of Phase 1. Therefore, the design will take into account features that will be added in later phases. At the end of this phase there will be a basic Ascel Bio administration interface that allows the management of the public forecasting webpage, the public forecasting webpage itself, and all the required hardware and software that support those.
Storage Subsystem:
Data ingestion must be flexible and allow the implementation of different data sources including; CVS files, XML Files, and REST APIs. Data ingestion pipeline must allow data to be directed to the text analysis subsystem. Deployed system will have a minimum storage capacity of 4 Terabytes. Additional storage capacity, up to 200 Terabytes, should be supported. Data stored in the system is immutable, i.e. data can be written, read and deleted, but not modified.
Publish forecasts and IDIS Trends to Public Forecasting Web Portal.
2.1.1 Deliverables
The primary deliverable of Phase 1 will be a production system deployed on Ascel Bio’s cloud infrastructure. As part of the deployment process, Ascel Bio personnel will be trained on how to use and administer the system. Because this is a full system deployment, the training will focus on DRI developed system components. Specific deliverables of this phase will include:
Detailed technical specifications. Functioning system deployed on Ascel Bio’s cloud infrastructure. Administration and Public Forecasting Portal documentation. Non-technical training on how to interact with the system. Technical training on how to add models to the system.
2.2 PHASE 2 REQUIREMENTS
Phase 2 will be built upon the software and hardware infrastructure developed in Phase 1 by adding additional features to the website and Text Analysis Subsystem. A Subscription Based Web Portal will be created that allows customers to purchase a subscription to the web portal for looking into various disease forecasts for the location or locations of subscriber’s interest.
Storage Subsystem:
Data at rest or in motion must be encrypted. Deployed system will have a minimum storage capacity of 10 Terabytes.
Subscription Based Web Portal:
User/group portal access control should be manageable from the administration interface. User/group should be able to purchase, edit, and delete subscriptions.
Administration Web Portal:
Customizable per customer/group data storage limits. Add/remove/group customer accounts. Access to web analytics platform used throughout site to track the usage. Add/remove informational articles displayed in the Subscription Based Web Portal.
2.2.1 Deliverables
The primary deliverable of Phase 2 is an updated system containing the new functionality that DRI will apply to Ascel Bio’s production system. Similar to Phase 1, Ascel Bio personnel will be trained on how to use the system. Training will focus on the features added in Phase 2. Specific deliverables of this phase will include:
Updated detailed technical specifications.
Deployed updated system on Ascel Bio’s cloud infrastructure. Updated Administration and Public Forecasting Portal documentation reflecting the interface changes. Subscription Based Web Portal documentation. Linkage to payment gateway, such as PayPal Non-technical training on how to interact with the system. Technical training on the new features.
2.3 ARCHITECTURE DESIGN
Ascel Bio requires a storage and analysis infrastructure that can meet both current and future needs. In general, analysis infrastructures based on relational databases (RDBMSs) are not easily scaled. Therefore, the Big Data software stack was specifically developed with scalability in mind. It allows the horizontal scaling of both storage and computational resources. Upon acceptance of this proposal, a more detailed specification will be developed that may result in modifying this architectural design.
The Ascel Bio website will be comprised of a Public Forecasting Portal, Administration Portal, and a Subscription Based Portal. Those portals will be built using a Content Management System (CMS). Only after a more detailed website specification has been created can the most appropriate CMS be selected. Like most web frameworks, the CMS requires a HTTP Server and a relational database to function.
The storage and analysis infrastructure is comprised of a text analysis, forecasting, and storage subsystem. Hadoop provides access to HDFS, a distributed fault tolerate file system. By replicating data across multiple servers, HDFS provides both horizontal storage scalability as well as fault tolerance. New machines can be added to a Hadoop cluster as needed. The system will then start replicating data to them. HDFS only provides a storage infrastructure, it does not provide the ability to read, write, and delete individual records, which is commonly done in relational databases. Apache HBase adds that functionality to Hadoop and the HDFS file system.
Apache Flume will handle transferring data to the storage and analysis system. Flume allows the creation of fault tolerant data pipelines. The text analysis subsystem requires incoming text data to be indexed and simultaneously stored. By creating a branched pipeline, the same data can be sent to Apache Solr for indexing and HBase for storage.
Apache Solr provides a Boolean query language that can be used to search indexed text. The new Exigence rule engine will take a list of queries and send them to Solr for processing. Those results will then be used by the Netica Bayesian Belief Network to generate IDIS Trends.
Most statistical languages like R, are not designed to process large datasets. The Spark framework allows a single statistical model run to be executed across multiple computing nodes. Distributing the model execution across multiple nodes will reduce the execution time. As the analyzed datasets grow, the number of computational nodes used for a single run can be increased.