Big Data – Does Size Matter?

Many things have been said / written about big data, e.g., definitions around a number of V’s (from three to seven) have been published, success stories, and surveys reporting a huge number of fails. What we often see is that companies tend to jealously look at the internet giants. Consequently, they try to follow them by flooding data lakes, facilitated by technology and beloved by IT.

But is it really the size that matters, or the data we should start with? For sure, driven by, e.g., digitization in all dimensions of our world, data from everything and everywhere is  “easily accessible. However, we are not the giants, we do not have that easy access, we do not have the competencies, … So we need to focus. Perhaps it is better to start rather small in an area, which we know quite well, where we can at least try to estimate the potential, and where we are able to evaluate whether we achieve the promises. Then, size is probably not the most important thing. It is more about the combination of capabilities, competencies, and processes aiming at extracting the right data and generate relevant business value out of it.

We describe our approach to strategically align big data projects with organizational goals in our recent white paper.

Advertisements

CESI2017 again accepted for ICSE

The fifth edition of CESI workshop is accepted as an ICSE workshop. It will take place 23rd of May 2017 in Buenos Aires, Argentina.

You will find all important information at CESI Website.

With regard to the topic, we had some discussions whether we should go back to the roots of empirical software engineering, in particular addressing empirical methods or whether we should address new, trendy topics such as Big Data. Finally, Big Data won (somehow). For this year, we are looking forward to papers addressing this topic, e.g., answering questions like:

  • how can we make use of big data in empirical software engineering?
  • if we talk about big data systems, how do we empirically evaluate them respectively do we need to tackle their development differently?
  • are there any empirical studies (case studies) about digitization available?
  • do we have big data from/about software development life cycle?

There are many more questions. Never the less, we also welcome papers presenting high-light empirical studies, lessons learned from conducting empirical studies in industry, as well as reports from industry about how they made use out of empirical studies?

Using IBM Watson for empirical software engineering research

We were wondering, whether it is possible to use new technologies to support the effort prone steps in empirical research. In particular, identifying relevant papers for systematic literature reviews, extracting information from them to decide whether they are relevant, and analyzing the relevant one with regard to the research questions is a effort consuming task. As we saw from the use of Watson in a medical environment, where they support decision making based on a thousands of scientific papers, we are currently attempting similar for empirical software engineering research. We started with a published systematic literature review, took the paper the authors considered relevant and fed Watson with it. <TB continued>

Cognitive Policing

Currently, we are overwhelmed with new terms such as Big Data, Smart Data, Predictive Policing, and many more. What do they mean in practice? From my perspective, they seem to be like umbrellas, trying to cover existing things under their roof, perhaps incorporating some new aspects. If we look into predictive policing, I would say, that it was done already for many years. Without evaluating it with regard to decision quality, it was narrow, that is, based on individuals’ experience and little data. The reason is that for most agencies technological support was not affordable. With decreasing cost of hardware a new interest in technologies from artificial intelligence arise. Consequently, predicting today is supported by the automatized analysis of tons of historical and real-time data. However, building appropriate prediction models and interpreting the results from the analysis of the collected data requires care and human reasoning. If we talk about cognitive policing, which is way of incorporating artificial intelligence approaches into policing processes, we think it is important to make sense-fully use of the data we can get grip on. Data have to be used for the well-being and the benefit of the society. Human shall be the final instance for decision making. Nevertheless, existing and perhaps even disruptive ways of policing will be more and more supported with new technologies. The technologies shall be seen as a means to an end not an end in itself.

Big Data Potential Analysis

Big Data – Wie nutzen Sie Ihre Chance?

 

Objekte der realen Welt werden durch den Digitalisierung und das Internet der Dinge auch informationstechnisch greifbar. Die Vernetzung von ehemals getrennten Systemen eröffnet völlig neue Möglichkeiten und die Menge an verfügbaren Daten und Informationen wächst zunehmend.

Big Data ist in aller Munde – doch wie kann Ihr Unternehmen daraus Nutzen ziehen?

Schnellere Fehleranalyse, bessere Einsichten in das Kundenverhalten, effizientere Verwaltung des Fuhrparks, vorausschauende Instandhaltung der Infrastruktur – hier kann Big Data in Verbindung mit den richtigen Datenanalysen ein Unternehmen nach vorne bringen.

Das Fraunhofer IESE unterstützt Sie dabei, Big Data als Chance für Ihre neuen und innovativen Geschäftsideen zu nutzen. Mit Ihnen zusammen, setzen wir Ihre Ideen in unseren Innovationslabs prototypisch um, bewerten Kosten und Nutzen und helfen Ihnen dabei, Ihr Produkt- und Dienstleistungsportfolio entsprechend zu erweitern. Wir sind neutral und unabhängig, kennen den Stand der Praxis und Forschung; besitzen ein erprobtes Methodenportfolio und sind in der Welt der Informationssysteme und der eingebetteten Systeme zu Hause.

Nutzen Sie die Chance für Ihr Unternehmen! Eine gute Gelegenheit dazu bietet das vom Fraunhofer IESE angebotene Seminar „Analysis of Big Data Potentials in Business“, vom 21. bis 22. Juli 2016 in Kaiserlautern. Nähere Information gibt es hier: http://www.iese.fraunhofer.de/de/seminare_training/analyse_des_potenzials_von_big_data.html

Big Data – Big Opportunities?

Besides what we are told by big data industry, what is the real benefit of Big Data for companies?

Does your company need Big Data? „

  • Your business is data-driven.
  • You need to combine and package information from a wide variety of data sources. „
  • You use BI solutions to help you make your decisions. „
  • You are no stranger to the three Vs (Data Volume, Velocity, and Variety).

If at least one of these statements apply to your organization, then the question of whether and how much you invest in Big Data is justified. The answer is not simple. On the one hand, there are the strategic benefits of Big Data; on the other hand, there are often major costs involved for hardward, software, training, etc. Why not systematically trading off costs and benefits and thus avoiding investments in the wrong cause like it is proposed here.

CESI an ICSE2016 Workshop

Fourth Intl. Workshop on Conducting Empirical Studies in Industry (CESI 2016) – An ICSE 2016 Workshop

Building on the results and momentum of the CESI workshops held during ICSE (2013, 2014, and 2015) this workshop is the follow-on workshop at ICSE 2016.

There are numerous challenges, and rewards, in conducting empirical studies in industry. Challenges emanate from the multi-faceted complexity: in products, systems and services; development, evolution and management processes; vendor-customer models and relationships; organizational settings; business contexts and dynamics; etc. Yet there are rewards for successful studies in terms of utility of the results in real-world situations.

A new element of CESI 2016 is that, beyond the research-methodological focus of previous workshops, we seek research contributions highlighting results of the empirical studies conducted in industry. We believe that this move will:

  1. further precipitate empirical research in the software engineering community, and
  2. engage industry participants from the point of view of the utility of the results emanating from empirical studies.

A long-term goal of the CESI workshop series is to create a vibrant research and practice community with focus on conducting quality empirical studies in industry hoping that their results will lead to improved software engineering practices, techniques, methods, processes, technologies, products/systems and services.

Download Call for Papers (pdf)

Technical Debt

Partly funded by the German Federal Ministry of Education and Research (01IS15008A), we are currently working on the topic of technical debt.

The project aims at providing a method and tool for the strategic planing of technical debt, in particular, in agile software projects.

More information will be available soon. THe project’s website can be found here (in German)