Archive for Business Intelligence

Applied Machine Learning: Optimizing Patient Care in Hospitals, Profitably

Why are so many industries exploring Machine Learning as a means of delivering innovation and value?  In my view, the technology speaks to a primitive urge – machine learning is like having a crystal ball, telling you what will happen next.  For a business, it can convey information about a customer before they introduce themselves.  On a personal level, when I consider what I would like to have information about in advance, the first thing that springs to mind is obvious: my health.  Am I about to get sick?  How can I improve my wellness and overall health?  If you are wearing a Fit-Bit right now, then you probably agree with me.

In my last blog, I shared some real-world examples for how the Hospitality Industry applies Machine Learning.  What about Health Care and Hospitals?  While hospitals have similar challenges, in that they accommodate guests who stay overnight, the objectives in health care are quite different, and changing rapidly.   The Affordable Care Act is driving new business models, incentivizing outcome based reimbursement as opposed to volume based reimbursement.  Unlike hotels, today’s hospitals are interested in ensuring their guests do not have to return, at least not in the short term.   They also need to manage costs in a way they have not been incentivized to do in the past.  Hospitals across the country are considering how predictive analytics can have a  meaningful impact on operations, leading to improved patient health and improving the bottom line.

The cost savings opportunity for health care providers is startling. Here are just three examples:

Reducing Hospital Readmissions – in 2014, Medicare fined 2,610 hospitals $428 million for having high hospital readmission rates. Leaving actual fines aside, industry analysts estimate that the overall cost of preventable readmissions approaches $25 Billion annually. As a result, hospital systems all over the nation are mobilizing to intervene, using ML to identify risk factors which are highly predictive of readmission. Carolinas Healthcare System, partnering with Microsoft, did just that. Using data from 200,000 patient-discharge records, they created a predictive model deliver customized discharge planning, saving the hospital system hundreds of thousands of dollars annually. Read the article in Healthcare IT News.

Clinical Variation Management – Mercy Hospital is partnering with Ayasdi to find the optimal care path for common surgical procedures. Using knee replacement as an example, the Clinical Variation Management software helps hospital administrators find clusters of patient outcomes, then enables the exploration of those clusters in order to correlate a metric (i.e. Length of Stay) with a certain regiment or activity. Watch this video to learn how Mercy Hospital saved $50 million by applying Machine Learning to an extremely common procedure.

Improving Population Health – Dartmouth Hitchcock, a healthcare network affiliated with Dartmouth University, is piloting a remote monitoring system for patients requiring chronic care. 6,000+ patients are permitting the hospital to collect biometric data (i.e. blood pressure, temperature, etc.) in order that nurses and health coaches can monitor their vital signs, and machines can predict good days and bad days. Quite the opposite of hospitality: Dartmouth Hitchcock is trying to keep the guests from needing to checking in! Read more about the Case Study from Microsoft.

Are machines taking over for physicians? No. The Patient – Physician relationship remains (and I think will always remain) central to the delivery of personal health care. However, it seems clear that ACA is providing significant rewards to health care providers who manage population risk better. Machines can help here: through data, machine learning can find risk “hiding in the data”.

Contact Blum Shapiro Consulting to learn more about how Azure Machine Learning can curtail hospital readmissions, identify variations in common clinical procedures and improve patient population health.

Drive Your Business with a Dashboard, not Intuition

Business Owners and Senior Management should be constantly monitoring the performance of their business.  Some managers may accomplish this by talking to the people who work for them.  Ultimately, however, management is called upon to make decisions, and they do so with the information they have available to them.   Often the biggest challenge is bringing all of the data together into something which can be understood.

Most business leaders want to base their decisions upon clear, reliable information.  But, this can be a challenge.  Information can be difficult to obtain in a timely manner.  The accuracy of the data contained in reports is sometimes questionable.  Management reports may be extremely detailed without providing critical metrics which are easy to locate.  So what do you do?

If you are a manager with some of these issues, I would urge you to clear your head for a moment and go for a spin in your car.  That’s right – turn on the ignition, be careful in the parking lot, put on your favorite music and go for a 30 minute drive.  While you are there, take a moment to appreciate the amazing information solution right in front of you – your car dashboard.

A car dashboard exhibits three powerful concepts.

Consolidation– my dashboard prominently shows me my speed, my engine temperature, the gasoline remaining in my tank, among other things.  Where does this information come from? Generally, I know – but not specifically, and frankly, I’m not that interested.  But to answer the question, the data feeding the three dials I mention above each comes from three distinct components, undoubtedly manufactured by three separate vendors.   Can you imagine if your car started behaving the way most IT departments do? “I’m sorry, but we can’t tell you how much gas you have left, because that information is in a different system.” I know what I would do. I’d get a new car.

Context – I’m not an auto mechanic. I have no idea what the correct oil pressure should be in my car, or at what temperature my engine overheats. Fortunately, my dashboard is color coded. When the dial goes Red, that is bad. I need to take corrective action to move the dial out of the Red, into the Yellow or Green.

Relevance – for some pieces of information, any kind of numeric measurement is not useful. I simply want to know if I should be concerned or take action. For example, the alert that tells me that I don’t have my seatbelt on (and the car is moving). Or the one that tells me that I need to check my engine.  Fortunately, the dashboard does not provide me with a diagnostic code – because I would not know how to interpret it.  But I do know how to respond to a Check Engine light.

Operating a business is not very different from operating an automobile.  In both cases, you are working with a sophisticated, complex system which has a lot of moving parts.  Further, you need to keep your eyes on the road in front of you.  Driving or operating your vehicle is not the time to be looking for information, and getting that information can mean having to take your car out of service, costing money.

Unlike automobiles, it’s up to you as the manager of the business to decide which metrics are Key Performance Indicators – the ones you need to keep your eye on to ensure optimal performance.  Therefore, recognize that you will need to address your organizational readiness to begin monitoring your business in the same manner.

  1. Define your strategy – have a clear understanding of what the critical metrics are, and remove data and information which is not important to monitor.
  2. Embrace Data Driven Decision making – organizations which have committed to data driven decision making take good care of their operational data, attending to data quality issues as they arise.
  3. Keep it Actionable – Your envisioned dashboards will need right-time access to operational data – depending upon what you are looking to accomplish, you may or may not need “real-time information”. What is important is that you have information in time to make a correction, and don’t spend time and money investing on technology capabilities which your organization does not need.
  4. Keep it simple – You’ll need easy to use tools, both for IT and Management – and there are plenty of them available today!

By implementing a dashboard, or set of dashboards for your business, you can expect to be able to apply better focus to the critical processes that drive your business, because “What gets measured gets done”.  You’ll make better decisions, and also make these good decisions in time to take impactful action.  Finally, dashboards can help you communicate your strategy and performance to your management team.

 

 

 

4 Reasons Power BI is Better than Tableau, Qlik Sense

iStock_000006412772XSmall

I enjoyed reading this article by Martin Heller in which the author analyzed three top Data Visualization products: Tableau, Qlik Sense and Power BI.  Heller does a nice job explaining how these products represent an evolution in BI, making real data insights attainable for non-IT business users.  Mr. Heller says that each of these make self-service BI “remarkably easy” for users throughout the organization, but that in his opinion Tableau stood out as the best of the three.

But if Heller’s analysis is correct, then his conclusion makes no sense.  He first cautions that none of these products is well suited for Enterprise Reporting;: an important point – if you are looking for financial reporting, you are looking in the wrong place.  He then details each product’s features at length, with a focus on ease of use and the breadth of visualizations available in each.  He notes that there is very little difference in mobile capabilities (the Android app for Power BI is an exception, slated for September release).  Finally, he concedes that Power BI offers the best value as compared to Qlik Sense, and significantly better value than Tableau.  My question to Martin is this:  if you cannot afford Tableau or Qlik Sense licenses, and therefore cannot truly democratize business insights from data in your organization – what difference does the rest make?

Here are four reasons why Power BI is the smart choice:

1. Better Value – I’m not talking about a few dollars here and there, I’m talking 4x – Power BI Professional is $10 per month per user, Tableau Online is $500 per year per user.  I’ll let you do the math.  Mr. Heller does state that Qlik Sense is less expensive than Tableau, but does not go into specifics.

2. SaaS model – who wants another server on-premises to publish BI reports and dashboards?  Business users love cloud services: they pay a simple monthly subscription for the insights and visualizations they need, and  they don’t need to ask IT for anything!  Further, they get accelerated product updates from the cloud – much faster than traditional IT shops can maintain.  The net result is that vendors inevitably achieve parity, and the question “which product has the best visualizations?” quickly becomes a zero sum game.

3. Q&A – Power BI is the only product with Natural Language Query capabilities.  You simply type a question (i.e. “What was our customer churn rate in the past 3 months?”) and Power BI selects a visualization for you to explore.  The visualization chosen may or may not have been created by the dashboard’s author.

4. Power Query – all of these tools offer simple connectors to databases, Hadoop, CSV files, cloud data providers, but neither Tableau nor Qlik Sense provides a data shaping tool with the capability of Power Query.  For real power data users, Power Query’s M Language has all the capability of an IT Pro’s data transformation package, with none of the IT headache.  Power Query can be used in Excel, and it can also be used in the Power BI Desktop application. In either case, analysts consume and shape data on the desktop, build reports and publish to the cloud.

But is Power BI easy to use?  Yes it is – I invite you to come see for yourself! Dashboard in a Day is a Power BI workshop which BlumShapiro Consulting is hosting in Hartford, CT on May 11, 2017. We are offering Dashboard in a Day sessions at No Cost to participants.

 

Adding User Configurations to an Analysis Server Cube Part 2

Part 2: Dynamic User Configurations

In Part 1 of this series, we hard-coded some MDX values into the cube. That approach works in that it produces the desired end result, but if the values need to change, a developer is needed to make it happen. What is needed is a way to persist the configuration values outside of the cube itself. In Part 2, we will create a configuration table to store the values. The structure is borrowed from that used by earlier versions of Integration Services:

Really, the only two fields absolutely required here are the Name and Value fields; the other two are added for administration and clarification. Next we’ll insert some fictional values into this configuration table:

Next, create a view that pivots the Configuration Name and Configuration Value fields with T-SQL code like this:


 

The dataset returned by this view will be a single row of data with one column for each Configuration named in the PIVOT section, and a static [DummyKey] value of -1.

 

Again, some purists may dislike my use of “SELECT * FROM …” in my view definition, but since I am limiting the columns returned via the ” . . . FOR ConfigurationName IN (…) . . .” statement of the PIVOT clause, there is not much chance of getting unneeded columns.

 

Next, add this view to the cube project Data Source View, then add it as a Measure Group to your cube. Delete the COUNT and the SUM(Dummy Key) measures that were added by the Measure Group wizard. Since there is only one row in the measure group’s base table, a SUM( ) aggregations for the configurations are fine. Lastly, since a Measure Group MUST be joined to at least one Dimension, on the Dimension Usage tab join the Configuration Measure Group to a dimension in your cube that meets the following criteria:

  1. The dimension has a member row with a key value of -1. (Data Warehouse designers typically add a -1 key as the “Unknown” member of the dimension table.)
  2. You will NOT be using the dimension in conjunction with the Configuration Values. This sounds rather counter intuitive based on cube design practices, but it is explained below.

Browsing the cube by any dimension OTHER than the one used to join the Configuration Measure Group will return the configuration measure values at every cube intersection. This is because you are actually selecting the [All] member of that one dimension, which includes the SUM of each Configuration Value. And since there is only one row at the [Unknown] member (Key = -1), the SUM at the [All] level is the one row. Browsing the cube INCLUDING the one dimension will show that the configuration values are ONLY available for the “Unknown” member, and not for any others. If your configuration values, whatever they represent, will NEVER be used with the dimension you have them joined to, then this is just fine. But if there is any possibility that the Configuration Measures would be needed for any and every dimension in the cube, then you need to do a little editing of the view. We’ll cover that in Part 3.

The advantage of this method over what was covered in Part 1 is that if the Configuration Values ever need to be changed, it is now simply a matter of changing a single value in a table and reprocessing the Measure Group instead of editing the cube design and redeploying the entire cube. To add additional configurations would involve the following:

  1. Add the entry in the table
  2. Edit the view to include the appropriate [ConfigurationName] in the PIVOT clause
  3. Refresh the Data Source View for the cube project
  4. Add a new measure to the Configuration Measure Group for the newly added column
  5. Deploy and process the cube

In Part 3, we will overcome the limitation of NOT being able to use the Configuration Measures for EVERY dimension.