Wednesday, February 27, 2019

Collect to Report with Google Cloud Platform

Collect, store, transform and present all with Google Cloud Platform.

This is my first attempt at Google Cloud Platform (GCP), with a little help on google apps scripts I was able to create this dashboard. I was earlier shy with GCP, thought it needs good coding skills. I am not a coder, but can do some simple scripts with Google search. After dabbling for couple of days in GCP - Apps script, BigQuery and Data studio, Mission Accomplished.

This is not a real time reporting, real time streaming can be done. BigQuery was not required, Google sheet is good enough. Google sheet can store up to 2 Million data cells. If your Google sheet grows, data operations are slower. BigQuery is high performance, primarily for online analytical processing (OLAP) and much more.

My objective was to find ' Where are 'Data Scientist' jobs in US?

This is my approach
  1. Collect data from web/API with Google script
  2. Save the data to sheet with Google script
  3. (Optional) Save data to Bigquery with Google script
  4. Report with Data studio (includes data transformation)
There are other better visualization tools, but I found overall 'Collect to Report' process very easy with GCP. Only 300 line of Apps script was used in this project.
(Job API was discontinued in Oct,2020)

To zoom, click zoom icon at bottom right corner 

Wednesday, February 22, 2017

FP&A Team,Processes and Tools

FP&A Summit by Innovation Enterprise was held in San Diego, Feb 14- 15th. Finance and FP&A Leaders shared their experiences and presented topics on key matters such as FP&A Processes, Recruitment, Team development and Industry insights.

Below are my key takeaways:

1.    Recruitment
a.    Good quality FP&A resources are hard to find. Start recruitment early, making it your #1 priority. Candidates needs to have a solid background in Accounting and in Finance, most importantly with a growth mindset.
b.    People follow leaders - Your past colleagues/subordinates will likely follow you. Ask them ‘How can you add value?’
c.    Quality – Initially, long hours of work are needed for accomplishment, but the quality of work is what matters most in the long-term. Instead of asking ‘Can you work long hours?’ ask them ‘Describe your day at work?’. This will help you to understand ‘What are they producing at the end of the day?’ Open ended questions will provide qualitative feedback on the candidate rather than closed questions that provide minimal response.

2.    Team Composition
a.    Teams should understand the industry, and also challenge the norms or ‘status-quo’ at the same time. Hire a good mix of people from varied industries to provide the most complete range of responses, opinions and solutions.

3.    Development, Goal Setting and Performance Review
a.    ‘What got you here, won't get you there’, identify the attributes which are required to achieve the agreed goals in the long-term.
b.    If you are managing international team, explain in detail what you need. This helps you to align the expectations, and reduces the cultural gap.

4.    Tools and Solution Implementation
a.    Most of them have a tool in place for FP&A, while others are still using spreadsheets.
b.    Although tools were positioned in Gartner ‘Leaders’ quadrant, some were not happy with them. Common complaints were - solution was inflexible; they are still dependent on the consultant to make changes. Tools do not matter, what matters is ‘solution flexibility, scalability, and self-reliant?’
c.    Some also had negative experiences with consultants, others loved them. Spend more time understanding the experience of the consultants, ‘Are consultants experienced with accounting/finance systems? Do they have similar industry implementation experience?’ to ascertain if they are the right fit together.

Hope this helps you. If you like it or have comments please post here.

Thursday, November 17, 2016

Power in BI = Power BI

Many times decisions are based upon existing Features. 'The Only Thing That Is Constant Is Change', we do not want to base a decision on existing product features. When comparing we take into consideration how other competitive products/services are evolving. Are they adding new features? How frequently?

Tableau and Qlik historically had lead over Power BI in Data Visualization. I am now convinced that Power BI has the edge over others, here is the Gartner BI and Analytics magic quadrant report which ranks Power BI as visionary.

This is why I believe Power BI has edge over other

1. Organizations uses on an average '13' apps (data sources) - Okta Research finding

Which applies, a business analyst will have at least 13 tables joins to make a meaningful story. He/she has to slog (to understand the tables/views/relationships) if there is content pack, to get data from source. There are plenty of app connectors available on all three tools- Power BI, Tableau and Qlik. Power BI has more app connectors than others, you can quickly and easily get the required data from apps. Microsoft has understood the  importance of participating in open source, as it has officially joined the Linux foundation today. Not only it has connections to existing Microsoft stack (cloud/on premise), but also other others - Salesforce, Google Analytics, Github and many others. 

2. We need to store the data before consuming it, need to save data in SQL and other formats

ETL will be more relevant than ever before, to make a sense from multiple data sources - a good (join) relationship is required. Tableau and Qlik can read files and other data source, but they lack the ETL layer. Microsoft stack (SQL, Azure SQL and others) provides tools to transform and save the data in SQL and other formats. The complex transformation can Microsoft stack, leaving the light transformation to Power BI. Power BI as a part of Microsoft family has tremendous advantage.

3. Price, Price, Price

Look no further. At $9.99/Month is the Power BI pro version, the least expensive of the all products.

There are features which Power BI lacks currently - as discussed here by Tableau. This features comparison will not be relevant in the future, Power BI will catch up quickly. For some scenarios maybe Power BI is not the best Data Visualization tool. As a generalist which most of us are, we will find Power BI to be flexible, powerful and affordable.

Friday, August 7, 2015

Moore's law is slowing down, organization with focus on cost advantage will have better chance to thrive

Is Moore's law about:

A. Smaller, faster transistors? OR
B. Economics of chip making?
My answer is 'A', until 'B' holds true.
"The culprits (for the slowing of Moore's Law) are the rapidly rising costs associated with technology development and the capital equipment needed to produce next-generation nodes" 
McKinsey identifies this trend in the semiconductor industry in this article, and many Industry experts agree with this trend.
With pressing economic challenges and slower technological breakthroughs, organizations are focusing on cutting costs and consolidating of the value chain (via Partnerships and/or Acquisitions) 
The challenge for decision-makers within an organization is:- 
  1. Do I have the right information from all functions within an organization to model and plan for the future?
Traditionally EPM vendors had 'whiteboard' models – i.e. models with some features for financial planning but none for Sales and HR. Implementation of these EPM suites took six months or more, with minimal interaction of the various models.  Host Analytics offers a true, unified Cloud EPM suite, where most of the features are right out of the box. Financial, Workforce, Sales and CAPEX planning are all part of Host Analytics EPM offering.

B. Rising costs - How do we manage or decrease its unfavorable effects
Breaking down costs through Activity-based costing (ABC) (or Segment/channels for Sales/Marketing) can help organizations to identify the activities (expensive production process, or segment with higher costs) which need further analysis. You can view this information in Host Analytics (Actual's loaded from ERP and other systems), identify the issues and plan accordingly
Host Analytics offers integration services to many ERP and other systems, and drill-down to transaction details is also available.

C. Rolling forecasts, reporting and what if scenarios - Volatility, Consolidations
With Host Analytics Cloud EPM Suite, organizations can create rolling forecasts and create the most realistic picture possible of the future. There is no complex setup to create a rolling forecast.
You don’t have to purchase a different system for close management or consolidations, its available within the Host Analytics suite. Yes, it can even offer multi-currency consolidations and complex intercompany and equity eliminations.
You can also run 'what if' scenarios for consolidations and view the impact on the Balance Sheet and Income Statements.

D. Modeling of non-Financial information
Host Analytics 'Modeling' module recognizes the need for modeling of non-Financial information and flexibility to model and plan using Microsoft Excel as a front-end. With Host Analytics Modeling, you can model e.g. sales leads/opportunities or inventory SKU planning and then link then to other models .

To find out more, register for a Free demo

The Host Analytics Future of Finance Tour Tours starts in San Francisco on August 12th, 2015 

Friday, April 3, 2015

Cloud CPM suits positioned to be next 'Leaders' in Gartner's 2015 report.

Cloud CPM suite providers -Host Analytics, Adaptive Insights, Tagetik and Anaplan are moving into 'Leader' quadrant. Gartner has published 'Magic Quadrant CPM 2015 Report' and identifies trends in CPM suites.

Current hosted solution providers - SAP , Oracle and IBM are current leaders in CPM solution offering. They are late to understand customers needs, as their focus was on their cash cows 'ERP'. Cloud CPM suite providers realized this opportunity, and now with high customer satisfaction, they will move in 'Leader' quadrant.

Three strengths which cloud CPM has are
  1. Pure-play 'Cloud' solutions- which translates to reduced costs per user and easy upgrades
  2. Mobile device accessibility- user friendly UI and web analytics
  3. Pre-built models for planning and consolidation, and function specific rich features. e.g. HR and Sales models.

Host Analytics, was the first pure-play cloud vendor. Host Analytics started offering its cloud CPM service in 2007. It maintains its 'visionary' vendor position. Host Analytics provides function specific rich features to HR and Sales models.

To make migration to Cloud, vendors offer migration paths and tools to move your models and data from legacy to cloud CPM suite.

If you have a need for cloud CPM solution, please contact us. We will be happy to assist you in moving away from 'Excel chaos' or expensive hosted solution.

Thursday, June 19, 2014

World Cup 2014 -Predictive Analytics Showdown

Are you in World Cup 2014 friendly pool? Get help from 'experts' on statistically predicting 'who would advance and win the 2014 World Cup'.

If your betting window is closed, enjoy the game.

The models being off-target and you lose, it's very likely you will forget the loss (Freakonomics, bad predictions are rarely punished). Statistical models always learn from past recent events and refine the variables. Hope you win next time.

Here are the collections of World Cup 2014 Data Analytics and Visualizations

Nate Silver Vs. Goldman Sachs Vs. Bloomberg

Remember Nate Silver, 2008 US General Elections Prediction?, He and his team at FiveThirtyEight have statistical models on various subjects- Politics, Economics, Science, Life and Sports.

I like the visualization of the model.

FiveThirtyEight explains the statistical models which calculates the probability of teams advancing to the World Cup 2014 final. This model uses ESPN's SPI (Soccer Power Index).

Bloomberg with BSports have a good visualization, you can add your predictions and visualize.

The third model is by Goldman Sachs (they have been doing this since last four World cups)

SAP vs SAS vs R

SAP Lumira has fun facts and profitability.

There is not much analysis by SAS team, except this blog.

R-Bloggers contributors have posted statistical predictions -one favoring Germany and other Brazil.

Statsinthewild has interesting stats and data visualization.
Data Visualization

My favorite tool to visualize and learn World cup data is Google Trends. It uses real-time search data and uses interesting visualization techniques. I LOVE IT !!!

Tableau uses ESPN SPI statistics

What's is the difference between Prediction and Forecasting?
Prediction is a estimate of a specific event in the future. Forecasting is identifying all possible outcomes with each having its own probability, the total probability of all events will be close to 1.

The world just woke up now to 'Predictive Analysis'. The below Google Trends graphs seems to agree

Subscribe to Updates