All posts by pujanalytics

A Quantified, Greener World

By Catalina Alvarado Pedro Olaya

TV shows like The Jetsons,  Star Trek or Futurama  and movies like Back to the Future, Ghost in the Shell  or The Matrix have created a collective imaginary of a future where our daily lives and routines unfold in entirely unforseen ways . The surprising (and scary) thing is that parts of this future are closer to us that ever and driving those changes are some amazing technological innovations like Internet of Things (IoT), 3D printingQuantified Self,   Drones (unmanned aerial vehicles) , Augmented RealitySixth Sense and (of course) analytics. Let’s take a glimpse

It is 2024, as you get up a few unexpected things happen. Your general health state was monitored over the night and will be constantly checked during the day by some devices, which with a display showing different symbols such as smiley faces, will tell you if your indicators are fine or if there is something to check anticipating any medical condition. As you leave home your shared-time driverless car service, using predictive models, will estimate which car should come and pick you up.  Also connected with the traffic system, your car will estimate precisely the right variables (speed, optimal road, etc) that would make the most efficient ride. In this future red lights are unnecessary, parking lots and total road usage are minimized, decreasing carbon emissions and traffic jams [1].

In your daily live the ability of things to communicate with each other and do analytics in the fly in an enhanced version of the internet (IoT) will make your live easier in many ways. For example when you are at the airport (supposing that they still exist) your waiting time due to delays or for luggage. It is expected that airplanes will be able to warn airports while fying for example about their exact ETA and precise information on the luggage it carries or predict if they need some preventing maintenance.

You  might realize that you have forgotten some important devices at home and that you need them as soon as possible then you might contact your prefered shared drone service to bring them, making the entire process faster, easier and greener.

In fact many of these technologies already exist or are in advanced prototyping stages.

Tech trends

Internet of Things (IoT)

It is based on the idea that the interconnectivity and interoperability among everyday objects will facilitate  the way we do things. The idea is that these objects can interact with their internal states and the external enviroment, allowing us in this way to monitor everything, boosting efficiency,  reducing waste, loss and cost. 

An example of an this concept is award winning Neurio. It helps your home appliances to become smarter by using WiFi, cloud technology and a smart pattern detection algorithms. It also helps you reduce your overall power consumption at home by managing your appliances.

It is believed that when IoT is applied to industrial processes, this will revolutionize many sectors and in this context is called Industrial Internet.

3D Printing

The hottest buzz word these days in techy world is probably 3D Printers. With these toys you can build anything ranging from a prothesis to a house and anything in between, for example an acoustic guitar, a SLR camera lens. This is becoming a real game changer in a lot of industries and a lot of regulation is expectd around it, for example in gun printing.

Drones

Use of drones has grown quickly. They are being used mostly for surveillance and transport. Amazon’s R&D Lab has prototyped some models for immediate delivery.  On the other hand the Office of the Information Commissioner of Queensland cites the advantages that drones bring to the military like they can stay many non-stop hours flying, they are much cheaper than usual military aircraft and they can be flown remotely so there is no danger for the flight crew.

From the point of view of analytics drones are very flexible when collecting information. They are able to capture diferent types of information from an almost unlimited range of locations within their scope.  However in Its capability for collecting information so efficiently is where the drones Achilles’ heel lies.  They do not distinguish between private and public information when capturing it.

Health, Fitness & Quantified Self 

These technologies consist of sensors that monitor and track vital signs, body movements as well as clinical data. We all know someone who uses the Nike Band, shoes sensors or the smartphone app that tracks their performance while exercising. This tracking trend has been expanding well beyond the world of fitness to include an entire family of devices that serve purpose known as quantified self (QS).

QS refers to the idea of personal analytics, by incorporating technology to the process of gathering personal data, might it be from your body, what you do, where you are, what you eat, etc. This trend has been possible by the lowering of costs of personal computing, sensors and other types of technologies that facilitates this tracking process.

 

[1]. In the  IBM Global parking survey the very dramatic effects of transportation in pur daily lives are presented. See also this blog post

RSS Feed

Advertisement

Analytics and Decision Making

By Andrea Espíndola Pedro Olaya

 

There has been a outburst of attempts to frame decision making with action in consulting environments. We present in this post some of these ideas. Our main references are The Mckinsey Engagement by Friga and personal communications with professor P. Noonan of Goizueta Business School.

In these methodologies typically you start with a main problem or key question on which you eighter want to make a decision or a recommendation to a decision maker. You start using  issue tree analysis (here represented and refered to as pyramid) and the MECE criteria, as explained in  The Minto Pyramid Principle or else, you decompose your problem/question in a pyramidal structure (see figures below) whose last level of  analysis consist of leaves that are “atomic” or minimal propositions. Ideas at any level of the pyramid must be neccesary and sufficient to logically imply the idea in the node at the level above.

Sin título

Most of the conclusions we derive from analytical models belong to the lowest (atomic) level, from the point of view of the issue pyramid.  This view explains the logical connection between analytics and optimal decision making.

Through analytics, it is possible to make relevant decisions from empirical data in order to give solution to a specific problem or to improve the conditions of a certain situation.

Analytics typically lies at the intersection of Minto Pyramid Principle and an invert pyramid that encodes the solution/recommendation to the question we started with as shown in the figure below. First the MPP identifies the main problem, its implications and sub problems. Then, using the information available and through a far and wide analysis it might generate solutions to the different questions, which in turn might give rise to different alternatives, allowing the decision maker to reach a tangible solution/decision.

Sin título

In real life, the situation is way more complex and part of this complexity is captured by the DELTA model introduced in  Analytics at Work. But this will be the topic of a later post.

RSS Feed

The World of Lean and Agile: Evolution, Map and Territory

By Andrea Espíndola Pedro Olaya

In the blog post of Nov 26 2013 by Maurice McGinley he explains Lean Start-up, Design Thinking & Agile Methodologies in terms of a particular feedback loop known in the philosophical theory of interpretation as the Hermeneutic Circle. The idea of this circle is explained in wikipedia as follows:

“The hermeneutic circle (Germanhermeneutischer Zirkel) describes the process of understanding a text hermeneutically. It refers to the idea that one’s understanding of the text as a whole is established by reference to the individual parts and one’s understanding of each individual part by reference to the whole. Neither the whole text nor any individual part can be understood without reference to one another, and hence, it is a circle. However, this circular character of interpretation does not make it impossible to interpret a text; rather, it stresses that the meaning of a text must be found within its cultural, historical, and literary context.”

Visually it takes this form:

Hermeneutic Web

The analogy is to a traveler equiped with a paper map that is navigating in an unknown territory to her. In order to reach her destination she must go back and forth between the map and the territory and constantly update her understanding of both, the model (map) and reality (territory).  This process is shown in the figure below.

Map Territory Web

Instead of going into the depths of interpretation theory we attempt to explain the core of these feedback loops using an analogy with the widely known Theory of Evolution.

Evolution

We see

RSS Feed

General Consulting Methodology

By Andrea Espíndola Pedro Olaya

In this post we highlight the most important aspects of general management consulting mainly to have a common vocabulary. Our main reference in management consulting is the book by Milan Kubr, Management Consulting :  A Guide to the Profession, edited by ILO. The copyright of the images below is theirs.

General objectives in consulting

Image

1. Reach the goals and objectives of the organization.

2. Solution of managerial problems

3. Discover and evaluate new oportunities.

4. Learning  Development

5. Changes Implementation.

kubr

What we propose is to blend this classical view of consulting with  agile methodologies, as explained in coming posts.

RSS Feed

What is Modularity? And What Does it Do?

By Andrea Espíndola Pedro Olaya

 

As mentioned before we will adopt a modular approach to both, analytics and consulting in analytics. We will now explore a little bit the concept of modularity.

Lets recall briefly the definition of modularity and the basic tenets of modularity theory as developed by  Carliss Y. Baldwin and Kim B. Clark.

Modularity  mesures the degree to which a system’s components may be  separated and recombined to create new functionality . Different parts of a modular system can be  worked on independently of one another. Modules can be connected  as long as they conformed to a  predetermined set of design rules. Our favorite childhood example of modularity is LEGO ®:

400px-Robot_(2950264120)

Modular design rules establish strict partitions of knowledge and  effort at the outset of a design process.

(Baldwin & Clark)

Imagen1

What does Modularity achieve?

1. On the one hand it helps to increase the manageability in complex environments by establishing well define boundaries and thus limiting the scope of interaction between individual elements or tasks, effectively diminishing the amount and range of cycling that occurs in production processes.

2. It gives an architecture in which concurrency is natural.

3. It adapts to environmental changes.

(Source: B&C)

We believe that by architecting the team executing an anlytics project as a modular structure we will benefit by all these gains.

RSS Feed

What is Our Approach to Analytics?

By Andrea Espíndola Pedro Olaya

Our approach to analytics is non traditional in the sense that we bring together frameworks that are normally separated and tries to fulfill the needs of our group at our university. The consulting team consists of faculty from diverse disciplines including computer science, industrial engineering, business administration and visual arts.

To take advantage from the diversity within the team, we will look at analytics from three different points of view.

A. The first, is the  holistic point of view of the enterprise. For us, analytics trascends data mining in the sense that it focuses not just in the creation of advanced models of Mathematical/stats/Artificial Intelligence /Natural Language Processing/Econometric flavors but it also includes a softer point of view. Such point of view  comprises concepts and frameworks such as :

Thomas Davenport’s DELTA, Enterprise Architecture, Human Resources , Strategic Dynamics/Systems Thinking, Change Management, etc.

And these are dealt with by professionals with competences in areas including business administration, design and industrial engineering.

B. Second, a technical perspective. This corresponds to the traditional data mining point of view. It comprises frameworks and areas such that:

CRISP DM, Model Creation and Management, Modular Ontology of Data Mining, Analytics Software and Hardware Architecture.

These tasks are usually carried out by professionals with a technical background such as Statisticians, Mathematicians, Computer Scientists, Econometricians, Economists, etc.

C. Finally, an information consumption point of view. This is intended to serve as a communication bridge between the model producers and the model users at all levels (strategic, tactical and operational) The corresponding frameworks for it are:

Graphic semiology, Information Visualization and Communication.

This point of view  is best undertaken by visual artists, business managers and communicators.

In an interdisciplinary environment like our university we have access to people with a  wide range of skills. This allows us to carefully divide the process of consulting in analytics in three different modules that best utilize these specific skills. We will explain these modules in forthcoming blog posts.

RSS Feed

Who are We Writing This Blog For?

By Andrea Espíndola Pedro Olaya

Welcome to our analytics blog. This is the first of a series of posts that will describe a methodology that we have put together to enable a group of high profile professors in Pontificia Universidad Javeriana to serve in a team for consulting in analytics. We will also from time to time include posts on a one-off basis, commenting on new developments or new tools, etc. 

In this post we will clarify who are we writting this blog for.

Intended Users:

  • First, people who would like to start or improve an existing consulting service in Analytics
  • Second, business people who are in the process of hiring a consultant and wants to have a reference on what are important criteria to use in this selection process
  • Last, but not least, academics interested in the following areas:
  1. Analytics
  2. Consulting
  3. Service Modularity

In order for to identify how relevant is a given  blog post to you we will use the following tags

U1 for consultants in analytics

U2 for people hiring a consultant in analytics

U3 for academics interested in analytics

We do hope you enjoy reading this blog. You are welcome to subscribe to our feed in the link below and please do comment in the form below .

The persons who wrote this post are:

co.linkedin.com/in/pedroolaya

co.linkedin.com/pub/andrea-espíndola-luna/57/ab3/86/

RSS Feed