Revolutionizing Healthcare with Data Analysis
Insurance companies are always trying to reduce the costs of healthcare for obvious reasons. It directly impacts their bottom line, keeps patients happy, and increases the efficiency of doctors. What’s not to like? Unfortunately, there are a large number of diseases that are nonspecific in nature and for which medical practitioners can only give vague advice such as “Exercise, stop smoking, and eat right”. One example is called metabolic syndrome.
But a new tool is emerging that can allow doctors to provide far more specific advice. Insurance companies have massive troves of patient data which can be mined to uncover trends and statistics that were previously hidden. For example, based on the patient’s age, ethnicity, gender, and other factors such as weight etc., it may be possible for healthcare practitioners to say “if you exercise for XYZ in minutes a day, you will reduce your chances of having a heart attack by X% in the next 10 years.” This advice is customized specifically for each patient and is backed up by hard data analysis.
Just another example of the value of Big Data.
Repulsing SQL Injection Attacks
An SQL injection attack is a particularly devastating attempt to compromise your database if your web application is not coded properly. For example, if you hardcode your SQL statements and append user input as a string instead of using parameterized queries, a clever attacker can construct the variables in such a way that it allows them to send their own SQL statements to the database instead of yours. It’s not as difficult as you may think.
But proper coding practices aside, it is important for you to have an additional application layer firewall sitting in front of your web app that screens incoming user inputs for malicious SQL injection statements. Hackers are becoming cleverer all the time and using a professional security system makes perfect sense from a business point of view.
Fake “Cloud” Products
These days, there is a tendency to term any online service as “cloud-based”. In reality, this term has very specific connotations in the industry including elasticity, “pay-per-use”, and self provisioning. But many services that advertise themselves as cloud-based failed on all three counts.
Fortunately, it seems that IT administrators in various organizations are fully aware of the shortcomings of these services and are researching the various products thoroughly before they make the decision. Still, they shouldn’t have to check basic things like this. But there is little that we can do to prevent this “cloud washing”.
Taking the Guesswork out of Data Analysis
Anyone who has tried to analyze a large set of data will tell you that it is very often a hit and miss approach. Sometimes you uncover important trends, and very often you waste your time searching for something that isn’t there. Even if you do find something, you’ll be unable to tell whether it’s a statistically significant or not.
Fortunately, it seems that a startup known as BeyondCore is aiming to make the entire process more automated and provide analysts with guidance as to possible trends without any human intervention. It does this by using complex models – Principal Component Analysis or PCA for example – that are able to identify anomalies based on certain variables as well as check whether or not those anomalies are significant. It then allows the user to examine these trends in greater detail.
While some may fear that this takes the “art” out of data analysis, my view is that it will instead refine the process which can benefit from the strengths of both man and machine.