================================================================================ METHODOLOGY DOCUMENTATION Big Data and AI in Business Operations: Google Trends Nowcasting of New Zealand Tourism Demand and AI Adoption in the Sector Dr Yuqian Zhang, 10 July 2026 ================================================================================ 1. DATA SOURCES ================================================================================ 1.1 NZ Monthly Overseas Visitor Arrivals Source: Stats NZ, International Travel (April 2026 release) URL: https://www.stats.govt.nz/information-releases/international-travel-april-2026/ Coverage: April 2016 to April 2026 (121 months) Frequency: Monthly Variable: Actual (unadjusted) overseas visitor arrivals Revision: Stats NZ may revise historical figures; figures used here are as published in the April 2026 release. The January 2023 figure was revised in the April 2026 release to correct a cruise ship passenger list error. Methodology: Arrivals counted from arrival cards (pre-November 2018) and Customs passport data matched with arrival cards (post- November 2018). Country of residence imputation applied August 2016 to October 2018. 1.2 Google Trends Search Indices Source: Google Trends (trends.google.com) Queries: "new zealand travel", "new zealand tourism", "new zealand holiday" Geography: New Zealand (country level) Date range: 1 January 2010 to 30 June 2026 Retrieved: 10 July 2026 Frequency: Monthly Scale: 0 to 100, relative to peak search share within the query window. A value of 50 means half the peak search share. Limitations: Google Trends uses a sample (not the full search universe). Repeated pulls of the same query will vary slightly. Adding or removing date ranges can rescale the entire series. Low-volume queries may be suppressed (reported as zero) due to privacy thresholds. The series measures searches within the NZ geography, not searches by overseas visitors. Source-market geographies would likely yield stronger correlations for the tourism application; the NZ geography was chosen as a transparent, reproducible baseline. 1.3 Tourism Satellite Account (Economic Contribution) Source: Stats NZ, Tourism Satellite Account (year ended March 2025) URL: https://www.stats.govt.nz/information-releases/tourism-satellite-account-year-ended-march-2025/ Coverage: Annual (year ended March), latest: March 2025 Variables: Total tourism expenditure, international expenditure, domestic expenditure, direct and total GDP contribution, direct and total employment. 1.4 NZ AI Adoption Statistics Sources: - AI Forum New Zealand / Kinetics (2025). State of AI in New Zealand. - MBIE (2025). New Zealand's Strategy for Artificial Intelligence. - NZIER / Spark (2024). AI Activity in New Zealand: SME Adoption. - Datacom survey data cited in MBIE (2025). Note: These are survey-based estimates. Definitions of AI vary across surveys, and figures should not be directly compared between surveys without noting definitional differences. 1.5 AI Tourism Case Studies Sources: Reseller News NZ (Air NZ), ChannelLife/NZ Herald (Sudima Hotels), Tourism NZ/GuideGeek press releases, NIWA, industry reports. All case study claims verified against primary sources as of 10 July 2026. Where primary reports used ranges (e.g., Sudima "10-15%"), these are reported as ranges rather than point estimates. 1.6 Academic Literature Sources: Publisher databases (Wiley, Elsevier, Taylor & Francis, Springer), Google Scholar, journal websites. Verification: DOIs verified by attempting resolution; journal rankings from ABS 2021 and ABDC 2022 lists. Date of verification: 10 July 2026. ================================================================================ 2. FILE DESCRIPTIONS ================================================================================ 2.1 nz_tourism_arrivals_google_trends.csv Monthly panel (121 rows) combining Stats NZ overseas visitor arrivals with Google Trends search indices. Columns: month YYYY-MM format visitor_arrivals Actual monthly overseas visitor arrivals (integer) gt_travel Google Trends "new zealand travel" index (0-100) gt_tourism Google Trends "new zealand tourism" index (0-100) gt_holiday Google Trends "new zealand holiday" index (0-100) gt_composite Simple average of the three indices 2.2 nz_tourism_arrivals_annual.csv Calendar-year totals of overseas visitor arrivals (2017 to 2026 partial). Columns: year Calendar year visitor_arrivals Total arrivals (integer; 2026 is Jan-Apr partial) peak_month Month with highest arrivals in that year peak_arrivals Arrivals in the peak month notes Context (e.g., COVID border closure, travel bubble) 2.3 nz_tourism_economic_contribution.csv Key economic metrics from the Tourism Satellite Account and MBIE. Columns: metric Description of the indicator value Numeric value unit Unit of measurement period Time period the value refers to source Primary source 2.4 nz_tourism_ai_adoption.csv AI adoption indicators for NZ business overall and the tourism sector. Columns: indicator Description of the indicator value Numeric or approximate value unit Unit (%, users, NZD, etc.) year Reference year source Primary source 2.5 nz_tourism_academic_literature.csv Peer-reviewed papers on Google Trends nowcasting and tourism demand forecasting. Columns: authors Author names year Publication year title Paper title journal Journal name volume Volume and issue pages Page range doi Digital Object Identifier abs_ranking ABS Academic Journal Guide 2021 rating abdc_ranking ABDC Journal Quality List 2022 rating topic Brief topic descriptor 2.6 nz_tourism_README_methodology.txt This file. ================================================================================ 3. ANALYSIS NOTES ================================================================================ 3.1 Correlation Analysis Pearson correlation coefficients computed between monthly visitor arrivals and Google Trends indices. Full sample (April 2016 to April 2026, n=121) and four sub-periods: Pre-COVID: April 2016 to February 2020 (n=47) During COVID: March 2020 to February 2022 (n=24) Recovery: March 2022 to April 2026 (n=50) Post-recovery: April 2023 to April 2026 (n=37) The full-sample correlation of r=0.71 between the "new zealand holiday" index and arrivals is mechanically inflated by the pandemic, because both search interest and arrivals fell near zero together and recovered together. The pre-COVID and post-recovery sub-period correlations are more informative about the relationship in normal operating conditions. 3.2 Composite Index The composite index is a simple arithmetic mean: composite = (gt_travel + gt_tourism + gt_holiday) / 3 More sophisticated composites (PCA-weighted, MIDAS-weighted) would likely improve performance and should be explored in follow-up research. 3.3 Charts All charts in the HTML research brief were generated with ECharts 5.5.0 using the data from these CSV files. Chart data values are embedded in the HTML and can be cross-referenced against the CSV files. 3.4 Reproducibility To reproduce the Google Trends data, visit: https://trends.google.com/trends/explore?date=2010-01-01%202026-06-30&geo=NZ and enter the three queries listed in section 1.2. Note that Google Trends indices are sampled, so values will differ slightly on each pull, but the patterns and correlations should be stable. To reproduce the arrivals data, download the latest International Travel release from Stats NZ or use Infoshare (Tourism > International Travel and Migration - ITM). The Python script that generated the CSV files is available alongside this dataset. ================================================================================ 4. LIMITATIONS ================================================================================ - Google Trends indices are relative and sampled. Small variations between pulls are expected and do not affect the substantive findings. - The NZ geography for Google Trends measures searches within New Zealand, not searches by prospective international visitors. Source-market geographies would be preferable for tourism forecasting and should be explored in follow-up research. - Correlations are not causal. Search interest may be driven by factors other than travel intent (e.g., general news or curiosity). - AI adoption survey data rely on self-reporting and varying definitions of "AI" across surveys. Figures from different sources should not be directly compared. - Case study data come from company disclosures and media reports, not independently audited sources. - The COVID-19 pandemic creates a structural break in the search-arrivals relationship. Models fitted on pre-COVID data may not transfer to post-COVID conditions without adjustment. - RECOVERY CHART NOTE: The January 2023 recovery percentage in Figure 5 uses 67.2% from the Stats NZ published "proportion of 2019" table. The computed value from the revised January 2023 arrival figure (266,432) would be 66.7%. Stats NZ revised the January 2023 arrival data in the April 2026 release to correct a cruise ship passenger list error, but the published percentage-of-2019 table was not updated after the revision. We use the official published table figure (67.2%) to remain consistent with the source document. Both values round to 67% when reported to the nearest whole percent. ================================================================================ 5. CITATION ================================================================================ If you use these data in academic work, please cite the accompanying research brief: Zhang, Y. (2026). Big Data and AI in Business Operations: Google Trends Nowcasting of New Zealand Tourism Demand and AI Adoption in the Sector. Research Brief, 10 July 2026. ================================================================================ Dr Yuqian Zhang | 10 July 2026 ================================================================================