Hello, I'm Zurab Kirkitadze
- Welcome to my personal website. I am zurab kirkitadze doctor of business administration. i finished all 3 steps of higher education in business administration. i also passed internship at national bank of georgia. i have published 8 scientific articles as in georgia and also in foreign countries. have work experience and i am active learner. i defended my phd dissertation in foreign direct investment, specifically effects of FDI strategies on host country economy (on the case of georgia). you can view my dissertation by opening following link: My dissertation
- synopsis of dissertation in georgian and english
About me
- I'm a self-taught data analyst with a passion for solving real-world problems using data. I specialize in turning raw data into actionable insights through data cleaning, analysis, and visualization. My projects showcase my ability to work with tools like Python, Pandas, Excel, and Power BI, and apply techniques such as data wrangling, exploratory data analysis (EDA), and dashboard creation. I've built interactive dashboards and worked on data-driven projects related to customer behavior, credit risk, and sales trends. You can explore my work on the Projects page. Currently, I'm seeking a remote data analyst role where I can contribute to data-driven decision-making and continue growing my skills in analytics, SQL, and machine learning.
- .
- I know all of the following skills of data analyst see here
My Education
- 2011-2020 Doctoral student of Tbilisi state university
- Faculty of Economics and Business, program of business administration
- module of international business
- see diploma
- Diploma Supplement
- 2011-2012 Intern
- National bank of Georgia
- Department of macroeconomics and statistics
- Monetary policy division
- see cetificate
- 2009-2011 Master of American University for humanities
- faculty of business administration (English master’s program of business administration)
- see diploma
- 2007-2009 Caucasus school of business for incomplete period
- Student on master’s degree course of business administration
- 2001-2005 Bachelor of American university for humanities
- Faculty of business administration (English Bachelor’s program of business administration)
- see diploma
- official transcript of records part 1
- official transcript of records part 2
- .
- .
My Skills
- Microsoft office programs Excel-very well, Word-well, Powerpoint-well
- here are my works with IF function
- Eviews 12-very well
- SPSS 23-very well
- Stata 12-very well
- R-Normal
- Python-Normal
- let’s break down ROI (Return on Investment): 🔹 Step-by-step: Find your total gain from the investment (money you made). Subtract the cost of the investment (money you spent). Divide that net profit by the cost. Multiply by 100 to get a percentage.
- Have performed inferential analysis as part of doctoral dissertation for example took sample and made conclusions about population
- Have performed gap analysis for example found out that current account deficit was not overcome in georgia and made recommendations how to overcome it with FDI for example that more resource seeking and efficiency seeking FDI should inflow into georgia to overcome this economic problem
- about proper usage of nominal and ordinal variables for training models
- Know google sheets and google forms. with latter you cen send the same form of questions to multiple recipients simultaneously. Google Sheets is basically Excel in the cloud: It runs in your browser (no installation needed); Multiple people can edit the same sheet in real time; It saves automatically to Google Drive and Some advanced Excel features (like certain macros or Power Query tools) are limited or work differently
- Banking regulations
- Apex ERP-Accounting Modulus-Normal see here
- Forecasting new product (this is product for which does not exist historic adoption data, one that has not been yet launched or exists at an early stage of adoption lifecycle and also one for there exists little to no market information about adoption of this product other than that there exists information about external (innovation) and internal (imitation) influences) adopters using Bass model see here
- Tableau Public-Normal
- Microsoft Power BI-normal
- Normal knowledge of Oris Accounting see here
- Know excel formulas NPV and IRR which can be used to make decision whether to accept or deny a project and also know the principle which should be used to make this decision using these criteria. Also know risk-adjusted return on capital (RAROC) that is a great metric to help you make decisions about whether to accept or reject a project. The idea behind RAROC is to measure the return of a project or investment relative to the risk taken on in executing that project.
- Know how to calculate entropy. From all available distributions that all satisfy given constraints for example if all of them have the same mean and standard deviation you choose one with maximum entropy. In the maximum entropy principle, we’re trying to pick the distribution that adds the least extra structure or assumptions beyond what we already know. This principle is used when you are uncertain what future real distribution will look like but in case you have strong basis to believe what will be future distribution you should choose it even if it does not have maximum entropy
- distribution generation with specific mean and variance and between some interval
- under all distributions defined by mean, variance and lower and upper bounds truncated normal distribution is the maximum entropy distribution
- the uniform distribution is indeed the maximum entropy distribution among all continuous distributions defined only by a lower bound and upper bound
- if only mean and variance are given, and the support is all real numbers, the normal distribution uniquely maximizes entropy.If we have [0,∞) with given mean → Exponential distribution is the maximum entropy distribution.
- The maximum entropy distribution for given[0,∞) and specified mean and log mean is a Gamma distribution.
- Discount rate calculation for NPV purposes. In georgia for risk free projects you can use monetary policy rate that is refinancing rate as risk free rate for discount rate
- BRD, FRD, and SRS
- Six Sigma
- system testing and UAT coordination
- Business process modelling, how jira works and UML diagram
- Some economic indicators
- credit scoring models
- Knowledge of Bash see here
- PESTLE and GAP analysis
- Principles of underwriting
- what is a prototype
- Have awareness of Azure ML concepts
- here calculations are made via excel formulas PMT(total payment per period), IPMT (interest) and PPMT (principal). on other same calculators calculations are made the same way
- Knowledge of chatgpt.com and solid ability to understand its answers that is i understand it answers very well
- Knowledge of financial instruments
- VBA-normal (know how to program code by searching in relevant sources) see here
- comprehensive knowledge of Bloomberg
- common collateral appraisal methods
- Model risk management
- tools that will help you make business decisions
- Decision tree to arrive at final decision incrementally
- Decision tree to arrive at final decision incrementally with longer path to final decision with another python code
- SQL-Normal see here
- MongoDB-Normal see here
- Data warehousing concepts
- A/B testing using two-sample test of proportions that is in its turn classic A/B test using STATA. This test is used to find out if difference is statistically significant between proportions and proportions in its turn reflect how many performed beneficial action of those who viewed website ad for example. Beneficial action shows for example how many purchased product of those who viewed ad. It measures how many people "converted" (e.g., purchased, signed up, clicked, etc.) out of those who were exposed (e.g., viewed the ad, visited the page). statistically significant difference means that it is real and not random so it has causes
- Comprehensive knowledge of recommendation systems in python, segmentation with clustering methods, cohort analysis, funnel analysis, ad-hoc analysis and retention modeling via such methods as cohort analysis, survival analysis and churn prediction models see here
- Data modeling for example to reduce data redundancy via normalized data modeling i create multiple tables that will be linked for example instead of mentioning specific customer details in order table every time he makes order i use customer ID in order table from customer table
- Data cleaning that is crucial step in data preparation that helps ensure the quality, consistency, and reliability of your dataset before analysis. see this on my personal projects page
- DAX-normal see here
- IFRS standards – basic level see here
- Version control Git and GitHub see here
- Basic knowledge of SnowFlake see here
- Basic knowledge of Apache spark including how to connect it with python see here
- Basic knowledge of Jira see here
- Attention to details-demonstrated strong attention to details by collecting only text relevant to dissertation title and chapters when working on PhD and master dissertations
- Agile methodology of project management see here
- Statistical analysis that is proven by the Diploma Supplement see here
- Pulling the data using an API see here
- Accounting, microeconomics and macroeconomics that is proven by bachelor's official transcript of records. you can also these lessons on my website of courses
- Knowledge of key performance indicators see here
- SWOT analysis see here
- Fuzzy logic model in matlab
- Know how to conduct sampling that will be representative of a population as i know random sampling when each individual in population has equal probability of being chosen and also know stratified sampling as the latter is good for representability because in this case we choose from each group for example when sampling from some school population we select from both boys and girls
- sample size calculator 1
- sample size calculator 2
- sample size calculator 3
- sample size calculator 4
- Excel formula to randomly choose between lower and upper limits is RANDBETWEEN
- Conducted qualitative and quantitative research/analysis for example in case of former i found out motivations of foreign investors for making foreign direct investments in Georgia and in case of latter i found out how much foreign direct investment inflowed in georgia according to years and also gathered primary data by conducting interviews with investors what was the reasons they made FDI in georgia and also gathered secondary data from the national statistics office of georgia which was placed on their website geostat.ge
- Google analytics including how to set it up on website see link here
- Debugging via Inspect Console that is a valid and important method for finding and fixing problems in a website
- Basic knowledge of HTML, javascript and CSS (i created 15 static websites)
- see some of them: (the companies on these websites are not real. They are solely created for the purpose of demonstration of these skills)
- truck website
- windows installation offer
- static website creation offer
- vegetable shop website
- other vegetable shop website
- this website
- this website with other alignment of pages
- fruit website
- website of flowers
- cafeteria website
- cake website
- machinery factory website
- second machinery factory website
- My cafe website
- My courses including econometrics, statistical programs, data science, credit risk analysis, data visualization tools, macroeconomics, accounting and GMAT
- to create a basic multi-page website using React with modern tooling
- to create a website with angular
- a simple React + TypeScript website starter
- how to run and view Redux + React app in browser step-by-step
- dynamic website that automatically learns the user’s preferences based on what they view or click — and then fetches content from a backend API accordingly
- unit testing with Jest
- knowledge of Chart.js
- dashboard in web-based environment created by chart.js
- knowledge of D3.js
- dashboard in web-based environment created by D3.js
- knowledge of Plotly.js
- dashboard in web-based environment created by Plotly.js
- knowledge of Apache ECharts
- dashboard in web-based environment created by Apache ECharts
- knowledge of highcharts.js
- company profit chart with highcharts.js
- knowledge of amCharts
- dashboard in web-based environment created by amCharts
- knowledge of Google Charts
- dashboard in web-based environment created by Google Charts
- To see all above charts in one place as one website pages
- creating charts using the Recharts library in React
- creating multiple charts simultaneously using the Recharts library in React
- creating multiple various charts simultaneously using the Recharts library in React
- to create a fully working Bar Chart using D3 inside a React component
- make a simple Bar chart using react-chartjs-2
- simple and clean Pie Chart using Victory in React
- how to create a line chart using Apache ECharts for React (echarts-for-react)
- simple working example of how to use Nivo in a React project
- All required for indexation of website
- some tools required for webmasters
- full Tailwind CSS webpage template
- About Figma
- All that is required for responsive design
- WCAG
- website performance optimization guide
- webpack
- CI/CD basics
- 2 free tools for webmasters
- Progressive web app
- Authentication flow
- All that can be performed with state management
- Little about java and mini golf game created with java
- Deployed websites on Neocities and Infinityfree
- Bayesian estimation of DSGE using python
- I am also versed in the following methods of credit risk assessment:(you can see these methods on the website of my courses, course 1)
- 1. Variance and consequently standard deviation
- 2. Copula method with R and Python
- 3. Fuzzy set method
- 4. Value at risk assessment by Monte Carlo simulation in Python
- 5. Various classification methods such as linear discriminant analysis with Python
- 6. Various classification methods in Python: AdaBoost Classifier, Decision Tree Classifier, Quadratic Discriminant Analysis, GaussianNB, Gaussian Process Classifier, KNeighbors Classifier, MLP Classifier, Random Forest Classifier and svm
- 7. Logistic regression and for small sample exact logistic regression with Eviews, SPSS or STATA
- 8. Fuzzy logic model in MATLAB and Python
- 9. Machine learning and deep learning methods, such as neural networks (such as Multilayer Perceptron and Radial Basis Function)
- 10. Survival analysis/Cox regression to evaluate credit risk, survival analysis in STATA and SPSS that is to find what time period passes from giving credit to customer to default event.
- 11. Sentiment analysis to find emotions in texts sent from borrowers to lenders
- 12. Quantile regression to find how default probability varies according to various levels of income for example.
- 13. Ensemble methods that is averaging results derived from various models to find output that is close to each model that is finding result that won’t be too different from almost practically correct result
- 14. Stress testing that is creation of econometric model to see how undesirable condition of borrower characteristics, industry-specific risks, and macroeconomic indicators will influence credit risk.
- 15. Time series forecasting methods such as ARIMA, Exponential smoothing, Seasonal decomposition, gradient boosting, MLP regression, Random forest, Support Vector Machines (SVM), Simple Moving Average.
- 16. BERT classifier and the Passive Aggressive classifier that are text classifiers in order to determine the veracity of promises made by borrowers.
- 17. Panel data regression to evaluate how borrowers behaved over time in order to evaluate behavior of credit portfolio that is to determine generally what factors stimulated default and what factors lessened default event.
- 18. Altman's Z-score, which predicts bankruptcy risk using financial ratios.
- 19. Reinforcement learning (RL), specifically Q-learning to decide whether to grant credit or not in python.
- 20. Credit Scoring Models
- 21. IFRS 9 and the implications of CECL (Current Expected Credit Loss)
- 22. Behavioral Scoring Models
- 23. Causal Inference: Understanding methods for establishing causality in my analyses, which can help in understanding the impact of interventions on default rates.
- 24. SHapley Additive exPlanations (SHAP) that helps explain the output of machine learning models by assigning each feature an importance value for a particular prediction.
- 25. Finding Nash equilibrium under context of game theory when players are lenders and borrowers
- 26. ARCH/GARCH to forecast variance/volatility for example when dependent variable is absolute percentage change in profit/income of borrower to forecast variance/volatility of profit/income and this way find out uncertainty in future period income/profit of borrower. conclusion is it that default probability is higher when future forecasted income/profit is characterized by higher variance/volatility. Also knowledge of ARCH/GARCH variations.
- 27. XGBoost for Binary Classification and fine-tuning the XGBoost model using Grid Search. Also LightGBM, Catboost and Adaboost
- 28. Autoencoders that is used for anomaly detection, which can be useful in identifying fraudulent or unusual credit behavior patterns.
- 29. Factor Analysis that is used for uncovering latent variables that might explain the default risk of a borrower
- 30. Recurrent Neural Networks (RNNs) or Long Short-Term Memory (LSTM) networks that could be beneficial for modeling sequential data (e.g., borrower payment history over time)
- 31. Mean reversion model to model the interest rate dynamics
- 32. Metrics such as Sharpe Ratio to assess risk-adjusted return in credit portfolios.
- 33. Financial analysis using financial coefficients from financial statements
- 34. Linear Programming (LP) for optimizing a credit portfolio with python
- 35. Bayesian Logistic Regression to predict whether borrower will default
- 36. Techniques like Generative Adversarial Networks (GANs) to generate synthetic data for training models when data is scarce.
- 37. Genetic algorithm for credit portfolio optimization
- 38. Knowledge of credit rating migration matrix including how to construct it.
- 39. Model Drift and Concept Drift: Techniques for monitoring the performance of credit risk models over time, detecting if and when they start to underperform due to changes in borrower behavior, market conditions, or economic factors.
- 40. Credit pricing model that is what interest rate to charge according to probability of default
- 41. Multilevel classifier in Python for credit risk prediction
- 42. Evaluation of market risk influence on credit risk
- 43. Default probability by Merton model
- 44. Calculation of expected loss assuming that borrower will use real options (expanding, contracting or abandoning). I can create binomial tree of present value without real options and also can create binomial tree of present value taking into consideration real options in excel and Python
- 45. Model validation that is testing how model performs on unseen data that is on validation data and backtesting using confusion matrix and accuracy and other measures
- 46. Basel regulations that give guidelines and requirements how to assess and manage credit risk
- 47. Credit portfolio diversification and concentration
- 48. what effect does GDP growth, interest rate and inflation have on credit risk
- 49. Assesing the liquidity of borrower
- 50. Assuring that incorrect model output does not discriminate towards good borrowers
- 51. GDPR
- 52. Kalman filter to derive more accurate estimate of default probability
- 53. LIME-Local Interpretable Model-agnostic Explanations, which is used to explain the predictions of any machine learning model, especially black-box models like Random Forests, XGBoost, or Neural Networks.
- 54. DoWhy that is a Python library for causal inference
- 55. AI Fairness 360 (IBM) toolkit
- 56. Quantum annealing with D-Wave
- 57. Expert system to evaluate whether credit risk is high
- see the several job requirements
- see the several job requirements
- see the several job requirements
- .
My Work Experience
- 2011 Jsc “charkhi-xazi”
Administrator of Web page
functions: Intermediation in creation of web page and running of web page
- 2009-2010 MLM Technologys
Sales manager
Functions: sale of products, Attraction of people to work in this company
- 2005-2007 Inspector, Tax Inspection of Tbilisi
Functions: training of tax payers in tax code of Georgia, prevention of tax law violation, drawing up a statement of the case in case of Tax law violation, Delivering tax requests to tax payers
- My pedagogical experience
have small pedagogical experience. I was performing seminars with students in business administration, specifically I was preparing tests for students, was checking their exams and was asking them questions about materials that they studied on lectures. My seminars were evaluated with 100 points from 100 points.
- see professor assistantship grade
- was according to this book
Contact Me
If you'd like to get in touch, feel free to drop a message below!
Email: kirki_1205@yahoo.com
Mobile: +995 595 51 54 09
Address: Tbilisi, Georgia
Research
- 2009-2011 Fund of Naumann, Strategic management activities and leadership style of Henry Ford. This research was conducted as part of masters thesis see diploma
- see cover of the book according to which this thesis was written
- see dissertation in PDF format
- My research interests mainly focus on the following areas:
- 1) Foreign Direct Investment (FDI) — Investigating the effects of FDI strategies on host country economies, with a particular focus on Georgia.
- 2) Credit Risk Assessment — Developing and applying advanced statistical methods for credit risk analysis, such as machine learning techniques and econometric modeling.
- 3) Data Science & Financial Modeling — Using machine learning algorithms and statistical models for financial forecasting and risk management.
- 4) Data analysis
- 5) Econometrics and machine learning
Publications
- 1. Zurab Kirkitadze, REGRESSION ANALYSIS OF FDI BY USING ECONOMETRIC TESTS,TSU P. Gugushvili Institute of Economics, Proceedings, P.418-424
- see article
- 2. Z. kirkitadze, Passed measures for attraction of FDI and provement of their consistency, Journal “New Economist” 2018, # 2-3, P.46-54
- see article
- 3. Z. Kirkitadze, Analysis of FDI determinants in Georgia, Journal “Economists”, 2018 #2, P.146-156
- see article
- 4. Z. Kirkitadze, POSITIVE EFFECTS OF FDI IN HOST COUNTRIES, Journal “Economics”, 2018, #12, p.36-42
- see article
- 5. Zurab Kirkitadze. Discussion of passed measures for attraction of FDI and affirmation of their consistency. World economy and international economic relations. International Scientific Collection.Kiev. GUL, 2018. ISBN 978-611-01-1079-2. P. 24-29.
- see article
- 6. Z. Kirkitadze, T. Shengelia, Y. Kozak. FDI motivation effects on Host countries. World economy and international economic relations. International Scientific Collection. Volume 3, Kiev. GUL, 2020. ISBN 978-611-01-1615-2. P. 15-20.
- see article
- 7. T. Shengelia, Z. kirkitadze, The Post-Coronavirus Economy of the World and Georgia, Journal “Economics and Business”, 2020, #4
- see article
- 8. T. Shengelia, Z. kirkitadze, HOW TO OVERCOME SOME IMPORTANT ECONOMIC IN GEORGIA WITH FDI. World economy and international economic relations. International Scientific Collection – Vol. 4. Kyiev, GUL 2021. ISBN 978-611-01-1865-1
- see article
- .
My personal projects
- links to my personal projects:
- 1. https://www.kaggle.com/code/zurabkirkitadze/notebookead71f46ca
- 2. https://www.kaggle.com/code/zurabkirkitadze/notebook4b5caef3e3
- 3. https://www.kaggle.com/code/zurabkirkitadze/notebook2f90928144
- 4. https://www.kaggle.com/code/zurabkirkitadze/notebookc408467bb1
- 5. https://www.kaggle.com/code/zurabkirkitadze/notebook87665e5a16
- 6. https://www.kaggle.com/code/zurabkirkitadze/notebook3d2179ac32
- 7. https://www.kaggle.com/code/zurabkirkitadze/notebookea2588c4a8
- 8. https://www.kaggle.com/code/zurabkirkitadze/notebook830f4eeb2b
- 9. https://www.kaggle.com/code/zurabkirkitadze/notebook8b08505787
- 10.https://www.kaggle.com/code/zurabkirkitadze/notebook155cfa68ec
- 11.https://www.kaggle.com/code/zurabkirkitadze/notebook0d47772431
- 12.here you can see my works related to sentiment analysis, outlier detection by autoencoder, text classification by bert and passive aggressive classifier, LIME and SHAP, AIF 360, recommendation system via q-learning and many other. when you open this link go to courses and then course 1
- 13. project of data cleaning on kaggle
- 14. not exactly project but some work of creation of dashboard
- 14. finding the best classifier automatically
- .
- .