Data Scientists Already know when Allergies will Suffer

Although the pandemic has had a negative impact in many areas, including the job market and many industrial sectors, there has also been a trend for data scientists to have more health projects than ever. Hungarian experts from Starschema, pioneers in data analysis, told us some interesting details about their work, where data can be put at the service of health.

Big data (an astonishing amount of data or information that arrives quickly and cannot be analyzed by traditional media) is one of the most advanced expressions of recent years that everyone knows has great potential, but few know how to take advantage of it. the exercise. Founded in 2006, the Hungarian Starschema, which deals with business intelligence and big data, is true artillery in this field: the company has already worked with clients such as Netflix, Disney, Facebook, and Tesla.

According to preliminary estimates, the company generated revenue of some HUF 5 billion last year, an increase of almost 20 percent.

The goal of data science is to develop and use algorithms that can be used to explore complex relationships using statistics, machine learning, and vast amounts of data processing techniques. Most of these correlations lead to new business insights: we can learn which marketing campaigns are worth communicating to each customer, what type of credit card we suspect is fraudulent, or what diseases certain genomes may be involved in. We have covered some applications in the past where data can be an advantage in business decisions:

One of the lessons of the pandemic and this year is that the pharmaceutical and healthcare industries are beginning to make better use than ever of the potential of data, and Starschema explains its expansion primarily through requests from large companies, including consultations from pharmaceutical giants. This is because pharmaceutical companies have been hit hard by market uncertainty about COVID-19. For business planning (improvement of production capacities, planning of sales activity, developments, etc.), it was not possible to start from the sales data of the previous year and the trends in 2020, so new models have implemented forecasting, backed by company technology and business intelligence services. necessary.

Last year, tamas volume According to the CEO, with the global acceleration of digitization, there is a growing demand from its customers to develop new digital products and data analysis services. “The market is very bubbly. The global analytics sector, currently estimated at $ 30 billion, is growing about 25 percent annually, and the momentum is likely to continue through the end of the decade, “Fulda said.

Allergens can also be satisfied with the data.

Most of these projects operate in complete secrecy since the type of data being processed makes it easy to infer what the competition is doing, so not all clients or projects can be named. The Star schema algorithm is also failing to apply one of the unnamed giant allergy predictors in the United States. The app can predict with 80 percent accuracy how the user will feel in the coming days. The prediction is based on a machine-learning algorithm and the data model that was created with it.

The Service is Mainly Based on three Data Sources:

  • pollen data tied to a specific geographic location,
  • weather data,
  • and user comments.

Based on this, the algorithm calculates what an allergic person can expect, and the more data you provide about how they feel, the more accurate the prediction will be over time. The app is currently only available in North America, but it plans to expand to Europe. A big blow to future developments is providing the user with information not only about their expected well-being but also about the pollen that they may be sensitive to based on their symptoms.

The company has also worked with Richter Gedeon Plc on a euro pharmaceutical research project. A problem area in euro pharmaceutical research is analyzing the shape of the so-called mitochondrial network in 3D micrographs. Mitochondria are basically the powerhouses of cells, these little organs that convert nutrients into the energy needed for our cells to function. Software that describes network architecture and machine learning can be used to automatically distinguish between healthy and damaged tissue. The methodology is an important step in drug discovery processes for which an automated solution does not yet exist. Previously, biologists analyzed cell structures and that was much slower,

He told us that data analysis is also a new way to automatically control the quality of MRIs. Windhager-Hell Esther, Chief Data Scientist at Star schema. Typically, clinicians cannot view the recording until weeks later, and if the quality of the recording is not good, for example, because the patient has moved, the subject will need to reschedule an appointment, which will take a long time. . The automated analysis tool determines immediately if the exam is suitable for diagnosis, and if not, another exam can be performed immediately.

Thanks to a partnership between Starschema and California-based tech giant Snowflake, a leader in cloud computing technologies, an integrated information system was made available on March 25 that currently provides one of the most comprehensive data services in the world. world and relies on trusted pandemic-related resources. Free and open data collection can be of great help to governments, major companies, and international organizations in making their extraordinary economic decisions and actions.

Its complexity is due to the fact that the developers have supplemented it with relevant information such as the geographic location of the diseases, detailed results of laboratory tests, the health infrastructure of the region, demographics (mean age, sex ratio, density population), or established epidemiological measures. Users can access data in a clean, analysis-ready format that they can easily and quickly integrate with their databases.

“Even a simple company can generate reports from a variety of metrics, from thousands of tables. Hence the need for data scientists to choose those of the reports that interest them. Find out where you can see if the result is not in line with past trends, leading to worse or better performance. If you do, causal exploration will also be necessary – you need to find out why that particular region is performing worse or better. It is very difficult to automate a process of this type, but the demand is enormous, so progress has been made in methodologies and research ”.

Data Scientists are Still in High Demand

Windhager-Hell says that a moderately skilled data scientist can get more queries on LinkedIn every day, and it takes a lot for professionals with that kind of knowledge. Although at-home data science courses are also available, there is no definitive college course as of now. The biggest problem, the expert said, is that universities cannot keep up with how quickly this technology is being developed. By the time the syllabus can be taught, it will be essentially out of date.

As long as there is no special training for data scientists or engineers, a science degree can be a good foundation, even with knowledge of mathematics, physics, programming, and economics. There are many courses related to data analysis on the net, even free ones, that can be used to gather a fairly high level of knowledge.

Post a Comment