Generative AI in Data Analytics (GenAI): How to Use It

Every day we read news about artificial intelligence (AI) — tools like ChatGPT, Meta.ai, Gemini are changing the way we work. So what is generational AI (GenAI), and how can you benefit as a data analyst? In this paper, we explain what GenAI is and how it is used at each stage of the OSEMN framework.

This post what is data analytics a continuation of its guide.

Difference between AI, machine learning and GenAI

AI (artificial intelligence) is the way machines can see tasks that humans usually perform — thinking, judging, combining facts. AI isn't new, it's been around for decades. Machine learning is a form of AI: the machine is given a lot of data and examples, from which it learns and builds models and predicts new results. Platforms like Meta, for example, learn from user behavior and personalize ads.

So why the current hype? The reason is generative AI, a special form of AI. Most AIs work on preset rules, and GenAI is able to learn from data and create new content — text, images, or even music — that has never been seen before. It uses deep learning: a large amount of information is provided to the machine, it learns samples, then when redirected with a question (prompt), it creates new, original material that mimics the input dataset.

LLM, ChatGPT, LLAMA and Gemini

These machines are actually multiple consolidated servers located in the cloud (cloud). This data and model hardware is called Large Language Model (LLM). The model that Meta BUILDS is called Llama (Large Language Model Meta.ai) and we communicate with it through the Meta.ai interface. OpenAI's model is GPT and is used with ChatGPT; Google's model is Gemini. With these tools, you talk in natural language — as if you were speaking to a friend.

Example: You ask Meta.ai to make a route for European travel with a friend who uses a wheel chair; it gives you a complete plan where it's available, and you save hours of exploring. GenAI is not just text — it can also create images, music, code and even video because it has been trained with relevant data.

How does GenAI work?

The process behind GenAI is as follows: first, the machine (think of it as a large computer) is given a gigantic amount of information. Through machine learning, the machine learns examples and features from this dataset. Then the machine is asked a question (prompt) and it creates a new, original material that usefully mimics the input dataset. GenAI is extremely useful in data analysis as it can rapidly process large volumes of data.

An important nuance: GenAI doesn't just replicate existing information-it creates new content that wasn't there before based on the examples it learned. It is this ability to «create» that distinguishes it from conventional AI and makes it so powerful for data analytics.

Different GenAI technologies and inputs

GenAI-nin girişləri (input) müxtəlif ola bilər: mətn, şəkil, səs, animasiya, audio, video və ya kod. Model kifayət qədər müxtəlif data ilə öyrədilibsə, çıxış da müxtəlif formatda ola bilər. Mətnlə öyrədilmiş model mətn yaradır; şəkillərlə öyrədilmiş model yeni şəkillər; kodla öyrədilmiş model isə proqram kodu yaradır. Bu çoxformatlı imkan GenAI-ni müxtəlif iş tapşırıqları üçün əlverişli edir — data analitiki üçün bu, həm mətn hesabatları, həm qrafiklər, həm də kod generasiyası deməkdir.

GenAI at every stage of OSEMN

GenAI accelerates every step of data analytics. Let's look at the OSEMN framework:

Obtain — Collect

GenAI is able to generate synthetic data that mimics an existing dataset when real data is missing. In healthcare, GenAI creates synthetic patient datasets that mimic real patient profiles but do not violate privacy — so researchers can perform powerful analysis with little real data.

Scrub — Clean

GenAI automates data cleansing: detects and corrects errors, outside values, and missing values based on learned patterns. Example: cleaning the scanner dataset in the supermarket — fixing anomalies like double scanned products or missing loyalty datasets.

Explore — Explore

The area where GenAI is most powerful. With advanced example recognition, it uncovers complex connections not seen by traditional methods and helps build graphs. Marketing analysts can explore a large consumer behavior dataset with GenAI and find hidden trends across different demographics.

Model — Modeling

GenAI automates feature engineering — defines variables in the dataset and proposes new model architectures. Financial analysts use GenAI to build credit scoring models and improve accuracy.

iNterpret — Commenting

GenAI creates intelligible explanation and visualization for complex models. Example: in urban planning, GenAI interprets traffic flow simulations and provides intuitive visualization and scenario analysis to non-scientist planners.

GenAI in predictive analytics

One of GenAI's most valuable applications is predictive analytics. Traditional models predict the future based on past datasets; GenAI strengthens this process in several ways. It accelerates feature engineering by automatically identifying variables in the dataset, proposing new model architectures, and even creating them itself based on dataset characteristics. As a result, the analyst is able to build more accurate forecast models in less time.

Example: financial analysts use GenAI to build and improve credit scoring models. GenAI not only offers new traits (such as costing patterns), but also tests different model architectures and improves forecast accuracy. This significantly shortens the work that previously took hours or days.

GenAI improves data quality

GenAI həm də data keyfiyyəti problemlərini həll etməyə kömək edir. Real data çatışmayanda o, mövcud datanı təqlid edən sintetik data yaradır — bu, xüsusən böyük datasetlərə ehtiyac olan və ya məxfilik həssas sahələrdə (məsələn, səhiyyə) faydalıdır. Eyni zamanda GenAI data təmizləməni avtomatlaşdırır: öyrənilmiş nümunələr əsasında səhvləri, kənar dəyərləri və çatışmayan dəyərləri aşkarlayıb düzəldir. Beləliklə, həm datanın həcmi, həm də keyfiyyəti artır — bu isə daha etibarlı analiz deməkdir.

GenAI's applications in business

GenAI-nin tətbiq sahələri çox genişdir. İnsan söhbətini təqlid edən çatbotlardan tutmuş, heyranedici vizual sənət və musiqi yaradan proqramlara qədər GenAI insan yaradıcılığının və səmərəliliyinin sərhədlərini genişləndirir. Bizneslər onu müştəri xidmətində (avtomatik cavablar), kontent yaradılmasında (mətn, şəkil, video), kod yazılmasında və məhsul dizaynında istifadə edir. Şirkətlər AI-ni inteqrasiya etdikcə, GenAI-ni anlamaq və istifadə etmək data analitiki üçün getdikcə daha vacib olur.

Concerns about AI

GenAI güclü olsa da, onunla bağlı narahatlıqları da nəzərə almaq lazımdır. Birincisi — dəqiqlik: GenAI bəzən inandırıcı görünən, amma yanlış məlumat («hallüsinasiya») yarada bilir, ona görə nəticələri həmişə yoxlamaq lazımdır. İkincisi — məxfilik və data təhlükəsizliyi: həssas datanı AI alətlərinə verməzdən əvvəl ehtiyatlı olmaq lazımdır. Üçüncüsü — qərəz (bias): model öyrədildiyi datadakı qərəzləri təkrarlaya bilər. Məsuliyyətli AI istifadəsi bu riskləri başa düşməyi və onları idarə etməyi tələb edir. Meta kimi şirkətlər «məsuliyyətli AI» prinsiplərini məhz bu səbəbdən vurğulayır.

How should GenAI benefit?

The key message is this: GenAI is all about thoughtfully and creatively applying technology to real problems, rather than technology itself. More and more data analytics tools are adding GenAI capabilities. Think about where AI can help in your work as a data analyst-but always check the results, because the outputs of GenAI are not always accurate. The best way is to try these free tools yourself.

Frequently asked questions (FAQs)

What is the difference between GenAI and ordinary AI?

Ordinary AI works with predetermined rules, while GenAI learns from data and creates new, original content (text, image, code).

What is LLM?

Large Language Model is a model that consists of multiple servers in the cloud, taught with a large text dataset. LLAMA, GPT, and gemini are examples.

Will GenAI replace the work of data analytics?

No, it's not replaceable-it's speeding up. The analyst still needs to ask the right question, test the results, and interpret them in a business context.

What is synthetic data?

An artificially created dataset that mimics a real dataset. It is used when real data is missing or privacy is sensitive (e.g. patient dataset).

Can I fully trust GenAI's performances?

No. A GenAI can sometimes be convincing, but can create false information. You should always check the results.

What is deep learning?

Machine learning is a pioneering form of learning; it is powerful in making complex connections between pieces of information. GenAI is exactly what deep learning uses.

What are the main concerns about AI?

Accuracy (hallucination), privacy, and data security, as well as repeating the bias of the model in the dataset. Responsible use requires managing these risks.

Which stage of GenAI OSEMN is most useful?

It helps at every stage, but it's especially powerful in research (explore) — it quickly discovers complex connections not seen by traditional methods.

How do I incorporate GenAI into their projects?

Practically speaking, there are several ways to incorporate GenAI into your workflow. When examining the dataset, you can use it to quickly extract insights from a large dataset. When you write a report, you can make the first draft an GenAI, and you can edit it. When the code is needed (e.g. a Python script), GenAI can write it. You can ask them for visualization ideas. The basic principle: Use GenAI as a helper, but let the final decision and verification stand with you.

More and more data analytics tools (Excel, Google Sheets, Tableau, etc.) are adding built-in GenAI capabilities. This means that in the near future, GenAI will be a natural part of the daily workflow, rather than a separate tool. That's why getting to know these tools now and learning to use them responsibly is an important advantage for a modern data analyst.

Conclusion: GenAI is a helper, not a substitute

GenAI accelerates and makes every stage of data analytics — from collecting to interpreting - more efficient. Creates synthetic data, automates cleaning, deepens research, helps model, and translates results into intelligible visualizations. But it doesn't replace data analytics. The analyst still needs to ask the right question, define the purpose, check GenAI's outputs, and interpret the results in a business context.

Ən yaxşı yanaşma GenAI-ni güclü bir köməkçi kimi görməkdir: o, təkrarlanan və vaxt aparan işləri üzərinə götürür, sən isə strateji düşüncəyə və qərar vermeyə fokuslanırsan. Bu texnologiya sürətlə inkişaf edir, ona görə onu indidən öyrənmək və məsuliyyətlə istifadə etmək müasir data analitiki üçün böyük üstünlükdür. Ən yaxşısı — bu pulsuz alətləri özün sınamaq və gündəlik işinə tətbiq etməkdir. Meta.ai, ChatGPT və Gemini pulsuzdur; onların gücünü əsl anlamaq üçün sınamaqdan başqa yol yoxdur.

Finished this guide! Now you know the entire OSEMN era of data analytics and modern GenAI capabilities.

👉 The main guide is: what is data analytics