GIS PORTFOLIO
NIKO WOOD
ABout me
Hi! I’m Niko Wood, I recently completed my B.A. in Human Geography with a minor in GIS at the University of British Columbia, along with a GIS Certificate through the University of Washington’s Professional & Continuing Education Program. Through this program, I was part of a student team partnering with Washington State’s Department of Ecology to map the state’s secondary textile resale industry and its equity implications. In my own time, I’m also examining the spatial logic of Israeli settlement placement and its relationship to nearby Palestinian towns in the West Bank.
I’ve always had a map in front of me, whether it was a placemat at the dinner table, a map on my wall, or an electronic talking globe, so finding GIS felt like finally discovering the right medium for how I see the world. I’m fascinated by the relationship between people and the landscapes they inhabit: how politics, culture, and economics shape the land, and how the land in turn shapes human decisions. GIS lets me visualize these interactions and uncover spatial patterns that might otherwise go unnoticed, turning curiosity about place into analysis that can inform better decisions.
Contact Info:
nikowood@earthlink.net
+1 (206) 931-7819
Current Project:
Washington State's Secondary Resale Economy
This collaborative GIS project, conducted in partnership with Washington State’s Department of Ecology, maps and analyzes the state’s resale and textile-disposal infrastructure to better understand how thrift stores, retail clothing outlets, repair services, and textile disposal bins support a circular economy. Our team of four is using advanced spatial methods, including overlay analysis, proximity and network analysis, and spatial regression, to answer questions about equity and accessibility of textile reuse, such as:
- Are textile disposal bins located near retail and resale stores, and what percentage of bins are within walking distance (e.g., 0.5 miles) of a resale establishment?
- What percentage of people are within a given distance of a resale establishment, how many of those residents are low-income, and what percentage of people are within that same distance of both a resale establishment and a textile disposal bin?
- Is accessibility to resale establishments correlated with demographic characteristics (such as median income, poverty rate, percent of households without vehicles, population density, RUCA codes, and SVI), and does this relationship vary across different parts of Washington?
We are also building a Circular Infrastructure Strength Index that scores each spatial unit based on the number and mix of clothing resale stores, clothing retail stores, repair services, and textile bins, and running network analyses to estimate how many people are within a reasonable driving time of resale establishments and what share of those residents own cars. Because this is a complex group project with an external sponsor, it also demands robust project management: we maintain formal risk, issue, and change logs, a shared schedule, and structured documentation (statements of work, meeting notes, and methods memos) to keep roles clear, track scope changes, and ensure our analyses and Experience Builder deliverables meet Ecology’s needs and deadlines.
Cities of California Election Results
This interactive web map explores the 2024 U.S. Presidential Election results across California, offering an analysis of how voters in different geographic and political contexts cast their ballots. At the county level, the map presents two distinct perspectives: one layer visualizing vote margins based on the total countywide results, and a second showing margins calculated exclusively from votes cast within incorporated municipalities, deliberately excluding unincorporated communities. A third county layer quantifies the vote shift between these two conditions, revealing whether the inclusion or exclusion of rural and unincorporated populations meaningfully changes a county’s apparent political leaning. California is an especially compelling case study because it contains some of the most Democratic urban centers in the country alongside rural counties that vote as Republican as anywhere in the Deep South. Visualizing this at the city level makes the geographic dimension of that divergence legible rather than abstract. Users are able to click on any individual city or county to see its own vote totals, percentages, and party margins.
University of Washington SCOPE
Participant Distribution
Remote Sensing Group Project
On the 2021 Heat Dome’s Effect on Agriculture in the Okanagan Valley
Replicated methodology on the paper
“The Legacy of the Home Owners’ Loan Corporation and the Political Ecology of Urban Trees and Urban Air Pollution in the United States.”
This project replicates a published GIS study on the enduring environmental impacts of historical race‑based housing discrimination, focusing on Seattle as a case study. Using 1930s Home Owners’ Loan Corporation (HOLC) grades, I examine how redlining relates to present‑day tree canopy and exposure to air pollution–related health risks (cancer and respiratory hazards).
Methodologically, I combined zonal statistics with spatial regression and clustering: zonal statistics were used to summarize tree canopy by HOLC grade for 2001 and 2011, and Geographically Weighted Regression (GWR) residuals were analyzed with hot spot analysis to identify statistically significant clusters of unusually high or low environmental health risk. The results show that Seattle does not neatly reinforce the common narrative that worse HOLC grades always correspond to worse contemporary environmental health. Significant clusters of high and low cancer and respiratory risk appear across A–C grades, with no strong, consistent concentration in formerly redlined (Grade D) areas. For tree canopy, Grade A neighborhoods had the highest mean canopy in 2001 (30.58%), with Grades B and C lower (18.89% and 13.15%), but the unexpected pattern is that Grade D had higher canopy (23.01% in 2001 and 20.72% in 2011) than both B and C.
These findings complicate the idea of a simple, linear relationship between redlining and environmental outcomes in Seattle. A plausible explanation is that some Grade D areas encompassed steep slopes, parks, or less‑developed land, where higher canopy persisted despite their discriminatory HOLC designation. While this does not negate broader national evidence linking redlining to environmental inequities, it underscores the importance of city‑specific context and land‑use history when interpreting the spatial legacy of redlining.
Storymap of Fictional Excerpts of Holy Mesopotamian Life
Least Cost Path
Scenario and Methodology
The problem being imagined here concerns the building of a new large greenhouse agricultural facility. This facility will use considerable amounts of electrical energy; this will require a feeder line to be built from an existing high voltage transformer station to the plant. Two paths shall be proposed with Optimal Path as Line: One that prioritizes building it through open land to keep it out of residents’ view, with the second option protecting any natural land. Distance Accumulation was also used to include the terrains topology, involving aspect and elevation for directional context, since going uphill costs more than flat terrain.
Examining Ecological Fallacy with Crime in Vancouver
Scenario and Methodology
This map was created to explore how scale affects the observed relations amongst crime reports and census variables (MAUP). I looked at how results varied when studying the same phenomena at two different scales
of aggregation–the census tract (CT) and census dissemination area (DA) scales. Regression analyses were used to model the relationships between reported residential break-and-enter crimes AND the total number of households AND the median total household income. That is, can knowing the median household income and how many households in total there are in an area enable us to predict how many residential B & E’s will be reported within that area? Furthermore, by examining the residuals, we can identify areas where the number of residential B & E’s are above or below what the model predicts and look for reasons behind those differences. Location Quotients were used to measure whether a variable was more present or more absent in an area than average across the landscape. Finally,
I examined the residuals of an aspatial regression analysis (GLR) to see if they are spatially autocorrelated, and then performed a geographically-weighted regression on the DA data to see if that reduces the spatial autocorrelation issues.
An Overview of Digital Terrain Modelling
and Multi-Criteria Analysis
Scenario
BC EnviroConsultants was recently awarded a contract from Environment Canada to produce a report on a rare plant species Spiranthes diluvialis that is known only from the southern Okanagan. The species has been observed to prefer low-slope, mid-high elevation areas with a southern aspect. A multi-criteria model is needed to highlight potential habitat and helps direct her search efforts. The goal is to identify ~700 hectares that are most likely to support habitat for this species.
Methodology
To create DEM rasters for aspect and slope, I first needed to interpolate elevation spot heights by generating a Triangulated Irregular Network (TIN). Afterward, I determined the factor weights using the Analytic Hierarchy Process (AHP), establishing that slope and elevation are equally important, while aspect holds significantly more weight. Armed with these weight values, I assigned them to each raster factor using the Suitability Modeler. Next, I focused on identifying habitable pixels for Spiranthes diluvialis. This involved transforming the elevation raster to isolate pixels that fall between 800 m and 1000 m. I also transformed the slope values to highlight low slopes as more valuable habitat compared to steep slopes. For the aspect analysis, I converted degree values into normalized scores ranging from 0 to 100, since 360 degrees is not linear. To pinpoint the best regions for Spiranthes diluvialis, I created another Suitability Model that included these characteristics but excluded the factor weights, allowing for a comparison between weighted and unweighted conclusions. Finally, I conducted a presence-absence test to categorize the pixels into three distinct groups, visualizing the differences between the weighted, unweighted, and combined data sets. It seems like weighted suitable locations tend to be West of the Okanagan basin while unweighted suitable locations are on the Eastern side. They both seem to congregate in the Southwest as well, causing the only combined locations to occur there. It’s also observable that the weighted locations are within more river valleys than unweighted. Ultimately, there’s 705 hectares of weighted suitable land, 681.25 hectares of unweighted suitable land, and 173.5 hectares of combined land.
Exploring Language Diversity
Between Edmonton and Vancouver
Scenario
The simple LDI maps reveal the actual diversity levels for each census tract, so a reader can see the full gradient from low to high linguistic diversity and make fine‑grained comparisons between neighborhoods. In contrast, the optimized hot spot maps distinguish meaningful “hot” and “cold” areas from random variation. The LDI maps answer “how diverse is this specific place,” while the hot spot maps answer where are the broader, statistically significant clusters of high or low diversity across each city.
Edmonton’s language diversity is normally distributed and practically unimodal, on the other hand Vancouver’s language diversity is bimodal and negatively skewed. This with the combination of the mean being below 0.5 while Vancouver’s is above, illustrates that Vancouver is a more diverse city in terms of people’s mother tongue. Though Vancouver is more linguistically diverse, there’s more than 3 neighborhoods in Edmonton that have higher LDI values than any Vancouver neighborhood. Greater Vancouver Area’s populated municipalities are in direct contrast with the further outskirts/farmland, including a small distinct buffer zone of no significant differences. Most of Edmonton’s neighborhoods contain no significant differences in mother tongue, with the center/more populated portion of the city having a lack of linguistic diversity and some of the suburbs ranking highly in language diversity, which is completely opposite to Vancouver.
Methodology
In the Language Diversity Index formula, represents the number of people speaking language in a census tract (and there are different languages potentially spoken in an area). That number is divided by the total population of that census tract; the fraction that results is then squared. Over each census tract, those numbers are summed up. The result is subtracted from one, and that is the LDI. Thus, the following line of code was used to populate the fields from census tract data.
Plain Text
langdata[‘LDI’] = 1 – (langdata.iloc[:,2:24]).divide(langdata[‘COL1’],axis=0).pow(2).sum(axis=1)Coccidioidomycosis Cases in California
How do environmental conditions and work around dry soil relate to county‑level coccidioidomycosis (Valley fever) incidence in California?
Seattle’s Constructional
Development from 2014 to 2024
Visualizes how the volume, cost, and type of building activity vary across Seattle ZIP codes over the past decade
For the following audiences:
Professional / policy audiences: City planners and policymakers who want to see where construction activity is concentrated and whether it aligns with long‑term growth and housing goals. Housing, transportation, and infrastructure planners interested in linking permit patterns with zoning, transit investments, and neighborhood change.
Industry and market audiences: Developers, contractors, and real‑estate analysts using permit volume, type, and cost to understand which ZIP codes are “hot” markets and what kinds of projects are driving demand. Trade businesses (materials suppliers, specialty contractors) looking for areas with sustained construction activity to focus outreach and services.
Public and civic audiences: Community organizations and advocacy groups tracking where large, high‑cost projects are clustering and how that relates to displacement or affordability concerns. Engaged residents, journalists, or neighborhood councils who want a synthesized view of a decade of construction without digging through raw permit datasets.
ENVIRONMENTAL ASSESSMENT OF GARIBALDI
AT SQUAMISH SKI RESORT PROPOSAL
Scenario
Northland Properties and Aquilini Investment Group of Vancouver engaged my services as a natural resource planner to evaluate the necessary changes to the proposed Garibaldi at Squamish Ski Resort project before the the project can progress. The BC Environmental Assessment Office concluded in 2010 that the project required additional environmental considerations and strategies to prevent detrimental impacts on vegetation, fish and wildlife habitat. The Resort Municipality of Whistler suggested that skiing may not be viable in the area due to elevation limitations.
Methodology
Ecologically sensitive protected areas including old-growth forest management zones, ungulate habitat areas, and riparian fish habitat corridors were incorporated into the analysis using the Add Surface Information, Buffer, and Clip tools. These layers were used to calculate the percentage of land requiring protection (where no development is permitted). As a result, potential ski resort locations were restricted to areas outside of these protected zones, above the 600-meter snowline, and within the defined project boundary. This approach minimizes environmental and economic impacts while maintaining ecological viability, ultimately identifying 39% of the project area as protected land. Nearly a third (32 %) of the project area is below 600 m elevation.