Apakah big data itu? Pertanyaan ini begitu sering kita jumpai mengingat istilah big data gencar terdengar saat ini. Artikel ini akan membahas lima hal penting tentang istilah ini sebagai panduan singkat Anda.

Big data berarti jumlah, karakter, atau simbol dimana operasi dilakukan oleh sebuah komputer. Big data kemungkinan disimpan dan ditransmisikan dalam bentuk sinyal elektrik dan direkam pada medium magnetik, optik atau mekanis.

Arti lain dari big data adalah data dengan ukuran yang sangat besar. Dengan kata lain, big data adalah istilah untuk menggambarkan koleksi data dalam volume sangat besar dan terus bertumbuh sangat cepat seiring dengan waktu.

Contoh Big Data

Berbicara apakah big data itu akan terlihat jelas dengan mengungkapkan contoh. Sampel yang paling nyata adalah Bursa Saham New York yang menghasilkan 1 terabyte data perdagangan baru setiap harinya. Atau hal yang paling sering kita sumbang adalah data pada media sosial. Statistik menyebutkan bahwa 500+ terabyte data baru dimasukkan ke dalam bank data Facebook setiap hari. Data tersebut umumnya berasal dari foto, video yang diunggah oleh pemakainya. Data lain berasal dari pertukaran pesan dan komentar pengguna.

Contoh lainnya adalah Jet engine yang bisa memproduksi 10+ terabyte data dalam 30 menit jam terbangnya. Dengan ribuan penerbangan setiap harinya, data yang dihasilkan dapat melampaui sekian banyak petabyte.

Mempunyai tiga jenis bentuk data

Big data terdiri dari tiga bentuk, yakni terstruktur, tidak terstruktur dan setengah terstruktur. Terstruktur Data apapun yang bisa disimpan, diakses dan diproses dalam format tetap disebut dengan data “terstruktur”. Selama periode tertentu, ahli ilmu komputer telah berhasil mengembangkan teknik mengolah data seperti ini lalu mengambil nilai darinya. Contoh data terstruktur adalah tabel gaji karyawan. Tidak terstruktur

Data apapun dalam bentuk atau struktur yang tidak dikenal digolongkan sebagai data tidak terstruktur. Selain mengandung jumlah super besar, data yang tidak terstruktur mempunyai tantangan dalam pemrosesan dan pengambilan nilai darinya. Contoh paling umum adalah sumber data heterogen yang mengandung perpaduan teks file sederhana, gambar, dan video. Contoh lainnya adalah hasil pencarian melalui Google.

Setengah terstruktur

Jenis ini mengandung data yang terstruktur dan tidak terstruktur. Kita dapat melihat data yang setengah terstruktur sebagai data terstruktur dalam bentuk tetapi sebenarnya tidak bisa dituangkan, misalnya, ke dalam sebuah tabel definisi terkait dengan relational Database Management System (DBMS). Contoh untuk jenis ini adalah data yang direpresentasikan dalam file XML.

Mengandung 4 Sifat

Pertanyaan mengenai apakah big data itu dapat terjawab secara gamblang melalui tiga sifat ini: volume, keragaman, kecepatan, dan variabilitas.

Volume di sini berarti jumlah data yang sangat banyak. Ukuran data memegang peran sangat penting dalam menentukan nilai dari data tersebut. Juga, agar mengetahui apakah data tertentu bisa disebut big data atau tidak, kita harus melihat volume data. Jadi, volume menjadi satu faktor yang patut dipertimbangkan saat membahas tentang big data.

Keragaman merujuk pada sumber yang heterogen dan sifat data, baik yang terstruktur atau pun tidak. Sebelumnya, spreadsheet dan basis data menjadi satu-satunya sumber data yang diperhitungkan oleh mayoritas aplikasi. Kini, data dalam bentuk surat elektronik, foto, video, PDF dan audio turut dipertimbangkan dalam aplikasi analisis. Keragaman data tidak terstruktur ini memberi andil terhadap hal penyimpanan, data mining dan analisa data.

Kecepatan berarti laju dalam menghasilkan data. Seberapa cepat data dihasilkan dan diproses untuk memenuhi permintaan menjadi faktor penentu potensi nyata di dalam data tersebut.

Dan yang terakhir adalah variabilitas yang merujuk pada inkonsistensi sebagaimana terpampang dalam data pada waktu tertentu sehingga menghalangi proses agar bisa menangani dan mengelola data secara efektif.

Data mining business analytics is the next big theme that we need to discuss after the definition of data mining. Drawing some samples for the topic clearly gets you to the point with specific goals in mind.

Before coming to the samples, let’s draw differences between data mining and data analytics. Data mining means the process of taking out information from large data sets. While data analytics takes one step further. The term refers to using the information then analyzing it for further usage. Companies or organizations utilize inspection, cleaning, transformation and modelling of the information during data analytics. At the end of the day, data analytics produce beneficial information for companies or organizations to make decisions.

Data analytics is part of the overall business intelligence processes besides data mining, artificial intelligence and machine learning. We need to understand the data mining process before coming to data mining business analytics.

Broad steps in data mining process

The first phase is business understanding. Make sure you know what overall objectives of your business that will lead to a data mining problem and a plan. The strong comprehension of the goal will also lead to a good data mining algorithm. For example, business understanding on finding out what customers buy the most.

The second one is data comprehension. The phase includes data gathering, getting insights, and studying subsets. For instance, the supermarket wishes to apply a rewards program that hope customers input their phone numbers when purchasing. This will allow the supermarket to access their shopping archives.

Data preparation serves as the third phase. The most significant stage encompasses computer-language data taking and its shifting into a form. From there, there is a modelling phase that brings together mathematical models to look for patterns in the data. The next phase is evaluation that also includes evaluation and review. The companies need to ensure the last step answers their business purposes. As the examples lay out, the final answer is knowing a list of products customers mostly purchase. The last stage is deployment which refers to making a report as the simplest form or formulating a repeatable data mining process to occur often.

The correlations between data mining process and business analytics

Data mining business analytics can take some important points from the data mining process. Let us reuse the supermarket as the clear example. The data mining business analytics, in general, includes the following stages:

Classification

At this stage, available data are analyzed then moved into discernible categories. From there, companies can take conclusions. In regard to the supermarket, the manager of the supermarket may utilize classification to group the types of groceries bought by the customers. For example, produce, meat, bakery, etc. The store owners can later learn more about the customers’ buying preferences.

Clustering

By essence, clustering looks similar to classification. Clustering, however, is less structured, providing simpler option for data mining. For example, the supermarket owner can categorize the products into food and non-food items.

Affiliation rules

Also known as tracking patterns, specifically based on linked variables. For instance, customers who buy specific items will likely to buy another second, related product. This will cause the store to know what will customers purchase next.

Regression analysis

Regression is used to identify the relationship between variables in a set. From there, the supermarket manager, for example, can plan and model a specific variable. Thus, the manager can come up with price points based on availability, consumer demand, and their rivalry.

Unusual pattern projection

Sometimes, the supermarket manager needs to study anomaly consumer behavior so that they can provide products when the unusual season strikes. For example, the manager can offer products during the first week in March that sees most male consumers. This paints an unusual picture that mostly welcomes female shoppers throughout the weeks in the month.

Data mining uses are widely-applied in at least five sectors in today’s business and service-based fields. The technological advance greatly assists companies and institutions that target their activities to satisfy customers.

Data mining uses come from unprecedented huge number of data that emerge every second in today’s internet-driven era. Data analysis stems from transactional statistic, comment, click, and view in the internet. Below highlights data mining uses in at least five segments:

1.    Healthcare

To analyse big data, IT expert at a hospital deploys multi-dimensional databases, machine learning, soft computing, data visualization and statistics. As the results are available, the hospital or health institutions can use them for service improvement.

For example, they can use the data for calculating number of patients in each category. The same holds true for formulating most proper health services for patients. Data mining can also be used for detecting fraud and abuse.

2.    Retail

In the retail industry, market basket plays a critical role. The area helps retailers, managers, business owners, and entrepreneurs to study buyer behaviour. In particular, this point focuses on market basket analysis.

This refers to a modelling technique that stems from a theory that says if we buy a specific group of goods then we are more likely to purchase another group of products or services.

They can apply the analysis result for laying out goods at stores. Moreover, they can apply differential analysis comparison of results among various stores to make their stores standing out.

3.    Education

In this sector, a new term emerges, Educational Data Mining. This deals with method development to find out knowledge that stems from educational environments. Educational managements expect the Educational Data Mining to provide the most proper student future learning attitude.

As such, the institutions can emphasize on what to teach and how to teach to their students. They can do research for innovative techniques to teach their pupils.

4.    Manufacturing

Manufacturing engineering is an important area within manufacturing that encompasses complicated techniques for multi-layered production processes. Data mining uses are applied for discovering patterns in the processes. It can be used for extracting the connection between product architecture, product portfolio, and customer needs data. Some outcomes from the techniques are cost prediction, product development period and relationships that bind some tasks.

5.    Customer Relationship Management