Big Data Big Law

Big data is changing our lives, influencing our way of thinking and solving problems, in order to adapt to the trend of the Times, the organization must learn to use data to speak, if sitting on a large number of data but helpless or indifferent, that and no data is the same. However, in the analysis of data, complete self-creation is undesirable, because there are a large number of experience can be followed and learned to save a lot of time and cost. Recently, Orionx.net's co-founder, Shahin Khan, published an article about his team's experience and discipline from the big data, Internet of things, and cloud computing markets.

• The cost of preserving data is lower than the cost of deleting the data. In addition, there are multiple backups.

Big data becomes possible because it retains enough data, so don't delete it anyway, because you don't know when it will be used and what legal risk it will be to delete it. The cost of preserving data is low, and if anything happens in the future, you can also find evidence from that data.

• Whatever the motivation to start collecting data, they will cause you to collect more data.

Most data collection efforts focus on ongoing activities, but once you know how to use the data, the willingness to get more data increases.

• Big Data Systems start smaller, but slowly grow larger, with no intermediate size.

There are few medium-sized big data systems, and once the idea of a project proves to be promising, it will quickly grow and hatch new projects as it grows rapidly.

• The data must flow to a place of value, taking into account the value of the context of the feature.

Unused data is an idle asset that is likely to result in a depreciation of value. If you consider big data as a workflow, you must flow data to the most valuable places.

• Never assume that you know what the cause is and what the impact is.

Most of the big data scenarios are valuable and worth the effort, but their causal relationships are complex, data incomplete, and user biases unavoidable.

• The ratio between the data and irrelevant data will gradually tend to zero.

There is a lot of data, but most of it is useless, and only a few are worth it. The more data collected, the more obvious this phenomenon, which means that irrelevant data is growing at a much higher rate than the relevant data.

• The ultimate goal of the analysis is to synthesize.

Synthesis is required after the analysis is complete, which requires the introduction of machine learning and cognitive algorithms.

• Time = Money = data.

Data is an asset, although it can appreciate, but most of the time as new data replaces old data, the value of historical data will be lower because its relevance will become worse. So you must know the "interest rate" of the data and know how fast it will depreciate.

• Large capacity-fast-variety-low value density vs. non-reproducible-unrelated-incomplete-incorrect.

The quality of data directly affects the quality of data mining.

• By giving you enough data, you can prove the "pros and cons" of things.

The greater the amount of data, the more difficult it is to find valuable information, the complexity of the data, the irrational motives and the ignorance may result in invalid conclusions, but on the other hand, the more data there is, the more evidence is available to support the hypothesis, and, in a completely scientific way, sometimes the approval rate is even closer to 100%.

• The conclusion of big data is often interesting but useless, but will eventually become effective and useful.

In the new media age, interesting but superficial content is much more than insightful content, value mining needs to have a deep understanding of the data, but it takes time.

• As data volumes grow, big data and high-performance computing (HPC) need to be combined.

If you have 200 rows of data, you can use a spreadsheet, but if you have 2 billion rows of data, you must use HPC. In addition, as the volume of data grows, mathematical and scientific knowledge is needed to transform the data into models.

Big Data Big Law