personal finance digital assistant - pfda
TRANSCRIPT
Personal Finance IntelligenceAnother approach to Personal Finance Applications by BeyondIT AS
Some facts about Personal Finance
This is US facts based on surveys, numbers Globally are even worst:• Just 32 percent of Americans keep a household budget and 10 percent Globally.[1]
• 30 percent of Americans prepare a long-term financial plan such as savings and investment goals, while 7.5 percent Globally. [1]
• Persons most likely to create future financial plans are those with at least some college education and those making $75,000 a year. [1]
• Americans spend 12-18 percent more when they use credit cards instead of cash, nearly same Globally. [2]
• 76 percent of Americans live paycheck to paycheck, up to 80% Globally. [3]
• Half of the American population has less than one month’s income saved for emergencies. (It is recommended that a family of four have at least $5,887 saved up or three times their monthly income at the poverty level). [4]
• 44 percent of US households are “liquid asset poor”. Meaning they have less than three months of savings. [4]
Source:
• [1] http://www.gallup.com/poll/162872/one-three-americans-prepare-detailed-household-budget.aspx• [2] http://www.nerdwallet.com/blog/tips/credit-cards-make-you-spend-more/• [3] http://money.cnn.com/2013/06/24/pf/emergency-savings/• [4] http://issuu.com/cfednews/docs/2014_scorecard_report?e=7117260/6529225
Existing Personal Finance solutions
• Existing solutions (software, apps etc..) are manually handled
• Manual input for every transaction you commit• Manual planning and budget• No Forecasting options
Reason why existing PF software does not help?
Usage of Software Household engagement Budget Planning Benefits from Software0
1
2
3
4
5
6
7
8
9
Good Neutral Bad
PFDA breakthrough: Our solution PFDA vs Standard PF software
Opinion on main features Standard PF vs PFDA• Automated Data Update-right from your bank transactions
• Automated Budgeting- based on predefined datasets of your segmentation
(ex: calculate the average of the segmentations you belong and sets that as budget)
• Automated Forecasting: based on Predictive analytics models and
ML algorithms
Service Categor
yStandar
d PF PFDA
Data Input 5% 95%
Budget 12% 88%Forecasti
ng 10% 90%
PFI: From Idea to the Final Product
Idea/PlanningTechnical and
Human Resources
Solution Architecture
Testing/QA and
Implementation
PFDA: The Idea in a nutshell
Personal Finance Digital Assistant -PFDA represents the tech solution for personal finance and budget planning embed into your digital bank account or as an application for both desktop and mobile users, life fed directly from your bank account(s). All you need is install your app, authenticate with your bank account and everything will integrate simultaneously. Further, you do not need to maintain your data input manually as data generated from your transactions will populate the app back-end service and you will be able to see in the application in real-time. The application will feature benchmarking against a pre-defined data set that is usually the data set which you belong (ex: average family with 3 children, with 60.000-80.000$ yearly income, from a selected region, etc..) to help you have clear picture of your economy against the population you match most. Machine learning algorithms will make the application act better by learning it from huge data sets coming from other customers that will be willing to share their transactions anonymously Based on the capabilities we mentioned, the application will be able to serve you fully automated personal finance assistance.
PFI: Planning and DesignTechnical solution and back-end development
Technical/human Resources to accomplish the projectBack End Technical Resources: Business Intelligence Servers for back-end/front-end solution, Load Balancer, Applications Servers
Human Resources:BI Devlopers, Data Scientists. UX Designers and App Developers
Solution Architecture
Back End: gathering data (ETL example)
Front End: user interface (dashboard example)
*Both designs are subject of changes based on the request and conditions. *Note that dashboards in the front-end are only if we want to analylize our finances, otherwise the goal of the project is to create a digital advisor that will alert you regarding your finances.
Back-End: Extracted bank transactionsHere are example of data generated from my bank account via ETL tool like SSIS, from csv to a table and from there starts the magic.
Staging of the data
• Data get stagged from data capture tables• MetaData and Data quality transformations• Defined mapping for the DataMart (small DW)• Defined etntities: Stagging for Dimensions,
Staging for Facts etc..• Aggregations, disctincts, counters and other
neccessary transformations
DW or Data Mart
• Define DW architecture: Kimball vs Inmon• Define Dimensions and Facts• Define Mapping• Defined data quality process: CheckSum,
Attributes etc..• Consider Big Data and other external data
OLAP solution on top of DW/DataMart
1. Better query performance2. Well established back-end service3. Set measures and important KPI’s4. Set flaging regarding benchmarking budget values
Front-End interface
1. Note: the main purpose of the project is having automated alert system embed in your bank account2. The extended GUI is just to get better picture of your finances3. You can do extended analytics yourself because its user friendly
QA and Beta Testing
1. Functional Testing 2. Security Testing 3. Exception Testing 4. Design/UX Testing
CONTACT US
Contact: Email: [email protected], [email protected]: +4794875183Web: www.beyondit.noTwitter: @beskotw , use: #DataForGood #PDFDASupport the idea on Kickstarter: https://www.kickstarter.com/projects/1399062786/personal-finance-digital-assistantSupport us on Crowdfunder: https://www.crowdfunder.com/beyondit
The man behind the Idea
Besim Ismaili, Data Scientist"A decade of experience in the field of Information Technology, where 6 years of experience in Business Intelligence, Data Analysis and Design, Data Warehousing, Data Modeling, Mathematical Modeling, Statistics and lately Data Science. • Expertise in Data Modeling, Data Analytics and Predictive Analytics SSAS, MDX and DMX• Experience in development of Data warehouse, Data Extraction, Transformation, and Loading ETL• Excellent record of accomplishment in developing and maintaining enterprise wide web based report systems and portals in Finance, Enterprise wide solutions of BI&Strategy applications• Two times finalist of DND (“Den Norske Dataforeningen”) best solution in Business Intelligence• Certified from the best Universities in the World including MIT, Harvard University, Stanford University, UC at Berkeley in relevant fields"'Contact: Email: [email protected], [email protected]: +4794875183Web: www.beyondit.no