Národohospodářská fakulta Vysoké školy ekonomické v Praze, nám. Winstona Churchilla 4, 130 67 Praha 3 tel.: +420 224 095 521, fax: +420 224 221 718, URL: http://nf.vse.cz
REVIEW OF THE BACHERLOR THESIS OPPONENT
Student’s name and surname:Konstantin Tsutskov
Thesis title: GOOGLE TRENDS AS AN ADVANCED WAY OF INDEXING CONSUMER SENTIMENT AND PREDICTING PRIVATE CONSUMPTION
Thesis opponent:Ing. Tomáš Lichard, M.A., Ph.D.
1 2 3 4
Assessment of the main aspects of the thesis:
1 Structure of the thesis (logical and systematic construction) ☒ ☐ ☐ ☐
2 Goal of the thesis (unambiguity of formulation, adequacy) ☒ ☐ ☐ ☐
3 Quality of the theoretical part ☐ ☒ ☐ ☐
4 Quality of the application/analytical/empirical part ☒ ☐ ☐ ☐
5 Adequacy of the methods to the goal of the thesis ☒ ☐ ☐ ☐
6 Sources (topicality, relevance, frequency) ☒ ☐ ☐ ☐
7 Quality of conclusions, fulfillment of the goal of the thesis ☒ ☐ ☐ ☐
8 Formal aspects (formatting, stylistics, references) ☒ ☐ ☐ ☐
Verbal assessment (describe the strongest and weakest aspects of the thesis; especially in case of a low assessment 3-4, concrete reproaches need to be expressed):
The thesis is in general very well written. There are some minor problems with style: sometimes the language is too flowery and metaphorical for a thesis. However, in comparison to some other theses, which treat theoretical and empirical parts as completely independent, with the theoretical part being „necessary evil“, the theoretical and empirical parts in this thesis are well tied together.
The topic is quite relevant. The search data is one solution to the problem of nowcasting.
There is a nice use of graphics and tables supporting the text.
There is however some slightly sloppy mathematical exposition in the description of the index construction, which leaves the reader a bit confused. For example, in Equation (4), the sum is ∑" 𝑤!
#$! . Not only is the meaning of indices not described in the text, but the mentioned sum does also not make much sense – I am guessing that it should be 𝑤# in the sum, otherwise we are just summing the same number k-x times? I guess k is the number of categories, and 𝑤! is the inverted volatility of category x, but in that case, the sum of inverted volatilities should be ∑" 𝑤#
{#$&} .
Complex assessment (it is necessary to state whether the thesis complies with the Methodological guidelines of the Faculty of Economics, University of Economics, Prague as concerns the quality of the content, the extent and formal requirements and whether the thesis is being recommended to the defense or not; the thesis can also be nominated to a special award, etc. ):
Despite some minor problems, the thesis is well written, with relevant sources properly cited. The empirical part seems to be competently done. Therefore, the thesis complies with the guidelines, and I recommend it to the defense.
Questions to the defense:
Why is nowcasting (e.g. using web searches) important? Why cannot we just wait for the official GDP figures? What are some other examples of leading indicators that are used?
In the thesis, you compare predictive power of the models by comparing their R2, which is alright for the thesis. But which test would you use to test the predictive power more formally?
Suggested grade: 1
Date: 8.6.2021 ...
Thesis opponent’s signature