In both cases you have my respect and support for trying to understand how HPC, AI, Big Data and Industry 4.0 can be used in your business.

I am honored and humbled that you selected me as part of your portfolio of experts for understanding how to leverage AI based technologies in your company. I am also expressing my gratitude to the StarwAI for supporting candidature as expert.

Please consider the following, prior contacting me:

• StairwAI is using a matchmaking tool to propose to each SME the list of AI Experts that fit the pilot’s scope and needs. Obviously you make the final decision on the Expert or Experts that would like to cooperate with. In addition, you free to select any Experts from the list of validated Experts, not proposed by the matchmaking tool. However, you may want to give a shot to the matchmaking tool first.
• You may want to drop me a line informally before moving forward, I will be happy to answer to your questions and take a virtual (or physical) coffee and discuss your ideas. In this way I may be able to understand if we are a good fit for each other and if I can really provide value to you.
• StairwAI is planning a second support program in 2023. Therefore if we will not have the possibility to work together in this round, do not despair; we will be able to do it at the next iteration.

Where to know more:

• AI4Europe
• StarwAI Project

Finally, let me thanks AI4EU for bring together AI stakeholders and resources, overcoming fragmentation issues and encouraging the acceleration of AI-based innovations. In particular for supporting SMEs that may be left behind due to their limited capability to acquire competences in this current market.

Happy to support the StarwAI Project and AI4EU in their mission of supporting SMEs in their strategies for #DataScience #AI #industry40 and #hpc .
Share on X

The post StairwAI and supporting SME in their AI based digital strategy: Open Letter appeared first on Francesco Lelli.

The benefit of getting familiar with Data Science Concepts

There are several benefits of getting familiar Data Science concepts. Here, I will list the most obvious:

• You will have another methodology in your portfolio of tools for conducting research.
• The certifications that you will obtain will be part of your CV and will be able to differentiate your profile.
• If you plan a a programming intensive and prototype driven approach for conducting your thesis (or research in general) you will be able to focus on the most interesting aspect from day one. This will prevent the unfortunate situation where you will be spending your time in catching up with your programming skills instead of really conducting research.

Courses Related to Data Science In DataCamp

This is a Python 101 course to perform Data Science:

• https://app.datacamp.com/learn/courses/introduction-to-data-science-in-python

This course is a bit more advanced compared to the previous one and focuses on Python specifically for data science:

• https://app.datacamp.com/learn/courses/intro-to-python-for-data-science

Here you learn to write your own custom functions, complete with multiple parameters and multiple return values, along with default arguments and variable-length arguments:

• https://app.datacamp.com/learn/courses/python-data-science-toolbox-part-1

This one complements the previous one and lets you practice doing a case study:

• https://app.datacamp.com/learn/courses/python-data-science-toolbox-part-2

With this you will learn how to improve the efficiency and readability of your code:

• https://app.datacamp.com/learn/courses/writing-efficient-python-code

And this one provides you with techniques for data cleaning and data (pre)elaboration:

• https://app.datacamp.com/learn/courses/cleaning-data-in-python

Finally, the following 2 courses will let you familiarize yourself with Tensorflow

• https://www.datacamp.com/tutorial/tensorflow-tutorial
• https://app.datacamp.com/learn/courses/introduction-to-tensorflow-in-python

If you like the idea of learning Data Science, consider sharing and spreading the voice:

A collection of #courses in DataCamp for getting familiar with #DataScience concepts using #Python . They are free if you are a #student. #DataScientist #pythonprogramming
Share on X

Bonus Course (in OOP):

This course is not strictly related to Data Science skills. However, the more the complexity of your code will grow the more you will need to organize it properly. If you are not familiar with OOP at all, this post is a good introduction to Object Oriented Programming.

• https://app.datacamp.com/learn/courses/object-oriented-programming-in-python

The post Using DataCamp for getting familiar with Data Science concepts appeared first on Francesco Lelli.

You may receive support from proper offices at the university, open source initiatives, as well as from the web. Here, I will share my favorite set of resources. Some are strictly related to the university where I am currently working, however, many institutions offer similar services.

Support for writing and related work:

• Universities have dedicated facilities (in Tilburg it is called Scriptorium) to help with writing assignments, such as a paper, a report, or your thesis. These facilities offer individual support for writing in both the language of your bachelor/master’s thesis (whether it is Dutch, French, Finnish, or something else) and English. You can also get help with searching for relevant background literature, consulting specific databases, and with questions about quoting and source referencing. Do not misunderstand me, they will not do the work for you but they can provide you with guidance and feedback.
• Libraries have a collection of best practices on how to reference sources in the proper manner and avoid plagiarism. Ref Cite is the one of the Library of Tilburg University: It is a self-study tutorial explaining the why and how of working with sources in academic writing.
• Libraries (here you can find the link to the one of TiU) are also the natural go-to place for retrieving material for your thesis. Nowadays your library is probably digital and offers your remote access via dedicated VPN. Nevertheless, your resources may also be available in paper format.

Support for Software and IT in General

• Some software is expensive, some is difficult to configure, and some has special licenses for students/researchers/faculties member. The IT support (here the link to the one in TiU) may help you in navigating this labyrinth and solve the various software (and hardware) related issues.
• Data are an important ingredient for your thesis. Your library (link again to the TiU one) may have access to particular packages (datasets) that are not available in a common setting.
• Data are also available in the web. Possible sources include the one mentioned in this post as well the datasets that I recently released.

IT Resources Offered for Free to Students

Being a student is (sometimes) an asset. Many companies and no-profit associations offer special support to students for experimenting and conducting their researches. Here there are two that I believe should be mentioned:

• Amazon Education: offers for free both classes as well as free computing resources for practicing cloud computing and related skills.
• GithHub education pack: offers an entire set of IT based services as well as the GithHub pro version for free to students that apply for this program.

Online Communities in the Web

Online communities may also be a source of support, mentorship, and exchange of ideas. Here are two that you should consider joining:

• StackeXchange (here you can see my profile) contains many communities that can be helpful for programming, configuring applications, as well as handling research in general.
• AskAcademia in Reddit is also a nice (virtual) place for exchanging opinions of academia-related issues.

Problems in Organizing Your Time and Delays

Deviation from what is originally expected is more common than you can imagine. Universities usually offer dedicated facilities for covering these problems. In TiU, this is the case of the Dean of Students. This is where you can find tailored support in case personal circumstances cause unforeseen problems for you. You can also develop a personal roadmap or working/study schedule.

Personal Coaching and Mental Health in General

Let’s face it, Covid times are not easy for anybody. In this respect, universities are organizing services that are supporting the mental health of students and staff. If you are experiencing difficulties in focusing, you may contact a specialized service (again, here you can find the one of TiU as an example) that can offer tailored support for your specific issue.

Other Online Resources for Your Thesis

A bit of self-promotion here. I am curating a few guides with the aim of offering quick (but effective) tips and suggestions for delivering a good thesis. In particular, I would suggest you to look at:

• A curated set of articles for improving the various aspects of your thesis
• A curated set of articles for moving from “Script Programming” to Object Oriented Programming.

Your #Thesis will benefit from all the help that it can get. Here you can find who can help you and how.
Share on X

This article (Your Thesis and Help: Where to Get It When You Need It?) is part of the miniseries on how to do a good thesis. You can see the full list of posts at the following link:

The post Your Thesis and Help: Where to Get It When You Need It? appeared first on Francesco Lelli.

Adopting healthy scientific practices is essential for conducting good quality research. This may be intuitive but it is hard to achieve and does not happen naturally. This video from the Royal Society introduces the problem:

The term “replication crisis” emphasizes that replication, a corner stone in science, is often considered of secondary importance. If we want to fight this phenomena it is important to conduct research using practices that promote the possibility to do that.

The open science movement not only offers you a way to get credited for your work (see this post on open access for more), but also tries to solve the problem of replicability. It takes into account several aspects at the level of researchers, scientific venues, funding organizations, etc.

In this article, we will focus on aspects related to the way research is conducted and the direct implications (and advantages) for your thesis.

I am assuming that you are not interested in falsifying your data or following any other intentional malpractices. In case you wonder, if you get caught, you could face serious consequences that range from immediate nullification of your thesis to expulsion and/or immediate termination of your internship.

As a single researcher (or small group of researchers) your attention should go towards offering to other scientists the possibility to access the data that you used for the analysis as well as your procedures. This video explain in a funny way why:

In other words, these are a set of good research practices that you should take into account:

• Carefully describe the procedures that you followed for obtaining the data.
• Make your data publicly available so that other scientists can check what you are doing. This includes your mentor and your second reader.
• Document each step of the process in order to be able to share this information with fellow researchers.
• Disclose the meaning of each field in the data.
• Disclose and share the programming code that you used for developing the data as well as the analysis.

Adopting these practices has several direct benefits:

• Your supervisors and/or reviewers will be positively impressed and predispose towards a higher grade.
• You will obtain a (very) good grade independently from your findings. In fact, your thesis is not about your findings: Showing you capability to follow the scientific method is way more important.
• Your work could, starting from your data, be published in a relatively easy manner as a tangible follow up of your thesis. Thus, differentiating your CV from other graduated students.
• A fellow student could decide to continue your work. Thus, you will be able to leave something tangible after you, instead of having your thesis completely forgotten.

In a nutshell, the bottom line is that good research practices will give you a better grade.

In this article, I limited myself to describing the scientific practices that a single individual should follow in pursuing his/her thesis. Adopting a more transparent approach clearly has several benefits and a direct implication on your grade. However, the problem of replicability is bigger than it may look like from this post. It is taken into serious consideration by the entire scientific community. In case you are interested in knowing more, this video gives you a general (and insightful) view on the problem:

#good #research practices like the one promoted by #OpenScience and #OpenData lead to better grades for your #Thesis as well as to a better #Science and more visibility of your work via #openaccess.
Share on X

This article (Adopting good scientific practices increase your visibility and the grade of your thesis) is part of the miniseries on how to do a good thesis, you can see the full list of posts at the following link:

The post Adopting good scientific practices increases your visibility and the grade of your thesis appeared first on Francesco Lelli.

Data come in all shapes and forms. If you are doing your thesis, you are in search of a proof of concept In other words, you are attempting to prove the validity of an idea or concept, not to produce an industry/ready solution . Therefore, most of the time, you do not need particularly large datasets. However, you want to be sure that they are of a sufficient quality. After all, you need to elaborate them without introducing too much noise.

What follows is a selected set of resources. It is not exhaustive and I am pretty sure that you will be able to find more by searching the Internet and asking your supervisor. However, this list is a good starting point in case you are trying to build a research/thesis proposal or you are stuck and in search of an idea.

Google datasets search:

• Link: https://datasetsearch.research.google.com/

You guessed right. Google has a dedicated search engine for datasets. It is freely available and index data that implement a particular schema.org format. Some of the data may be behind paywalls. However, academics and students usually can contact the data provider in order to get access to them or to a selected portion that is sufficient for their study.

EU Open Data Portal:

• Link: https://data.europa.eu/euodp/en/data/

Once again, you guessed right. The European Union is committed to foster a strong data transparency policy. At the time of writing this article, I was able to find 15399 different datasets that are freely available. They cover a large variety of topics and are related to all the various arguments under the jurisdiction of the EU.

Yahoo Finance:

• Link: https://finance.yahoo.com/
• Link to the help center: https://help.yahoo.com/kb/download-historical-data-yahoo-finance-sln2311.html

There are plenty of financial DB in the web. Yahoo finance is the most known and straightforward. In case you are looking for daily quotations of various financial assets, this is the place for you. The second link will explain how to download historical data.

More on financial data:

• Link Thomson Reuters API: https://customers.reuters.com/developer/apis_tech.aspx

Some programming knowledge is required. An Application Programming Interface (API) gives you the possibility to access to the data of a website in a programmatic way. Reuters has a dataset that includes financial news and press releases. All you have to do is to write a few lines of code for accessing this information.

Kaggle.com, a community approach:

• Link: https://www.kaggle.com/datasets

Kaggle offers for its members the possibility to access and share data. In addition, it offers also a set of demo code for accessing and manipulating the data. The community is data science driven. However, some of the datasets do not require particular programming skills.

Get your data by yourself:

• link: https://www.programmableweb.com/

There is an ever growing amount of websites that offer the possibility to access them programmatically. Programmableweb is a directory that tries to list them. In addition, it lists links to the proper resources for using them. Note that use the APIs will require some programming. However, you may end up creating original datasets and this is already a valuable outcome of your thesis.

Where to get data for your thesis? In summary:

I hope that by now you realized that datasets are far from been a sparse resource. This collection or resources can serve you as the starting point for finding what you are looking for or for refining/developing your research idea.

Do you need #data? Here you can find a collection of resources for finding your dream datasets. #DataScience #BigData #OpenData
Share on X

This article (Where to get Data: a collection of resources for your thesis) is part of the miniseries on how to do a good thesis, you can see the full list of posts at the following link:

The post Where to get Data: a collection of resources for your thesis appeared first on Francesco Lelli.

If you are enrolled in the other 50% of the universities, your (candidate) supervisor can’t assign you a thesis directly but you need to find and propose a topic that will be evaluated and, if properly presented, it will become your thesis.  

Both these approaches (with thesis proposal or without thesis proposal) have merits and problems. I am not going to take a particular side. At the same time, if you have to write a research proposal I will give you a few suggestions in order to “make it fly” so that you will not have to write it over and over again.

You can consider a thesis proposal as a 2-3 pages document that describes what could be your argument of the thesis. Usually, there is not a strict format or a proposed template. However, you may want to ask to your thesis coordinator or your student office in order to be sure. 

All you have to do is to provide a small description of your idea and propose something that is feasible and original.  Nothing more and nothing less.

A good thesis proposal has 4 elements:

• The idea
• The methodology
• The tool
• The data

Let’s see them one at the time.

The idea:

Capture in 2 sentences what you would like to investigate. Maybe you would like to use a neural network for evaluating the possibility of predicting the financial distress of a company. Alternatively, you would like to investigate how companies adopt a particular cloud computing solution. Just use your fantasy and your curiosity. At the same time start thinking to who could be a possible supervisor that will be interested in this particular topic.

The methodology:

How would you like to investigate the idea that you are proposing? Sometimes is easy, think about the first example that I made in the previous paragraph. “Neural network” is part of the idea itself and clearly suggest a methodology. At the same time, some ideas require additional specification. In the second case, for example, you are not specifying how you intend to evaluate the adoption of a cloud computing solution. If that is the case, you have to make explicit what you intend to do. Maybe you will survey executives in several companies, maybe you will conduct interviews, maybe you will look at financial data and so on.

The tool:

Now that idea is clear and the methodology is explicit is about doing what you are proposing. In other words, what tool will you be using for your investigation? Continuing with the first example, maybe you will use a particular library (for example TensorFlow) and you will build a prototype using a programming language (for example JAVA or Python). In the second case, assuming that you will conduct a survey, you will use excel or R for analyze the data and some web tool for distributing questionnaire.

It does not really matter what you plan to use as long as is feasible, you can picture yourself doing what you are proposing for the next months and is a good fit your particular case of investigation (i.e. your idea).

The Data:

Now that you have a tool, what data will you use for validating or invalidating the claims that you will be making in your thesis? In the first example, you may want to consider some financial data-set that is publicly (or not) available. Alternatively, maybe you can see if researchers use a particular benchmark and model your thesis around that idea. In the second example, if you will be survey practitioners, you will actually build the data set that you plan to analyze.

As you can see these 4 arguments are interlinked and influence each other. (i) Can you build a thesis around a neural network if you do not have data set? (ii) Can you analyze a survey if you do not have access to people that are actually working at that particular topic? (iii) Can you build a prototype if you are not familiar with programming? The more these 4 aspects will be clear the more they will marge in a flawless manner. In other words the more you want to show that you ideas is original and feasible the more you have to seek a way to integrate these 4 aspects in the proposal. This is the secret of a well written proposal!

One more tip for your thesis proposal:

There is only one remaining aspect that we need to consider. Your work do not exist in isolation but is part of a particular discipline and you must show that you are aware of what others are doing in a similar context. As general rule, you have to propose something that is relatively original. However, at the same time, it does not have to be a “game changer”. The chances that someone else is doing (or did) something similar is quite high and you want to cite a few scientifically relevant works of other researcher in the field. At the same time, you want to specify where your work will be similar and where it will differ. For example, you may want to use a similar approach (Neural Network) on a different set of data that is richer or more precise/imprecise compared to the one that other authors are using. Alternatively, maybe the problem of cloud adoption has been used only via interview and you are proposing a similar investigation but with a different methodology (a survey for example).

For this investigation, google scholar is what you need. Just find 4-8 scientific articles that are relevant and spend 1 or 2 paragraphs discussing similarities or differences.

In summary, a thesis proposal is a combination of an idea, a methodology, few tools and data. What you are proposing in these 4 categories should integrate smoothly. At the same time you need to spend some time to investigate if other authors have done something similar. You need to do that for both, use their work as a source of inspiration and ensure some novelty of your approach.

How to write a #thesis #proposal or a #research proposal: a few tips and suggestions
Share on X

This article (How to write a thesis proposal or a research proposal: a few tips) is part of the miniseries on how to do a good thesis, you can see the full list of posts at the following links:

The post How to write a thesis proposal or a research proposal: a few tips appeared first on Francesco Lelli.